Merge pull request #1095 from LvyuanW/lvyuan/dev

fix(admin/accounts): reset edit modal state on reopen

.gitattributes

@@ -4,6 +4,13 @@ backend/migrations/*.sql text eol=lf
 # Go 源代码文件
 *.go text eol=lf
 
+# 前端 源代码文件
+*.ts text eol=lf
+*.tsx text eol=lf
+*.js text eol=lf
+*.jsx text eol=lf
+*.vue text eol=lf
+
 # Shell 脚本
 *.sh text eol=lf
 

.goreleaser.simple.yaml

@@ -47,6 +47,8 @@ dockers:
       - "ghcr.io/{{ .Env.GITHUB_REPO_OWNER_LOWER }}/sub2api:latest"
     dockerfile: Dockerfile.goreleaser
     use: buildx
+    extra_files:
+      - deploy/docker-entrypoint.sh
     build_flag_templates:
       - "--platform=linux/amd64"
       - "--label=org.opencontainers.image.version={{ .Version }}"

.goreleaser.yaml

@@ -63,6 +63,8 @@ dockers:
       - "{{ .Env.DOCKERHUB_USERNAME }}/sub2api:{{ .Version }}-amd64"
     dockerfile: Dockerfile.goreleaser
     use: buildx
+    extra_files:
+      - deploy/docker-entrypoint.sh
     build_flag_templates:
       - "--platform=linux/amd64"
       - "--label=org.opencontainers.image.version={{ .Version }}"
@@ -76,6 +78,8 @@ dockers:
       - "{{ .Env.DOCKERHUB_USERNAME }}/sub2api:{{ .Version }}-arm64"
     dockerfile: Dockerfile.goreleaser
     use: buildx
+    extra_files:
+      - deploy/docker-entrypoint.sh
     build_flag_templates:
       - "--platform=linux/arm64"
       - "--label=org.opencontainers.image.version={{ .Version }}"
@@ -89,6 +93,8 @@ dockers:
       - "ghcr.io/{{ .Env.GITHUB_REPO_OWNER_LOWER }}/sub2api:{{ .Version }}-amd64"
     dockerfile: Dockerfile.goreleaser
     use: buildx
+    extra_files:
+      - deploy/docker-entrypoint.sh
     build_flag_templates:
       - "--platform=linux/amd64"
       - "--label=org.opencontainers.image.version={{ .Version }}"
@@ -102,6 +108,8 @@ dockers:
       - "ghcr.io/{{ .Env.GITHUB_REPO_OWNER_LOWER }}/sub2api:{{ .Version }}-arm64"
     dockerfile: Dockerfile.goreleaser
     use: buildx
+    extra_files:
+      - deploy/docker-entrypoint.sh
     build_flag_templates:
       - "--platform=linux/arm64"
       - "--label=org.opencontainers.image.version={{ .Version }}"

Dockerfile

@@ -9,6 +9,7 @@
 ARG NODE_IMAGE=node:24-alpine
 ARG GOLANG_IMAGE=golang:1.26.1-alpine
 ARG ALPINE_IMAGE=alpine:3.21
+ARG POSTGRES_IMAGE=postgres:18-alpine
 ARG GOPROXY=https://goproxy.cn,direct
 ARG GOSUMDB=sum.golang.google.cn
 
@@ -73,7 +74,12 @@ RUN VERSION_VALUE="${VERSION}" && \
     ./cmd/server
 
 # -----------------------------------------------------------------------------
-# Stage 3: Final Runtime Image
+# Stage 3: PostgreSQL Client (version-matched with docker-compose)
+# -----------------------------------------------------------------------------
+FROM ${POSTGRES_IMAGE} AS pg-client
+
+# -----------------------------------------------------------------------------
+# Stage 4: Final Runtime Image
 # -----------------------------------------------------------------------------
 FROM ${ALPINE_IMAGE}
 
@@ -86,8 +92,21 @@ LABEL org.opencontainers.image.source="https://github.com/Wei-Shaw/sub2api"
 RUN apk add --no-cache \
     ca-certificates \
     tzdata \
+    su-exec \
+    libpq \
+    zstd-libs \
+    lz4-libs \
+    krb5-libs \
+    libldap \
+    libedit \
     && rm -rf /var/cache/apk/*
 
+# Copy pg_dump and psql from the same postgres image used in docker-compose
+# This ensures version consistency between backup tools and the database server
+COPY --from=pg-client /usr/local/bin/pg_dump /usr/local/bin/pg_dump
+COPY --from=pg-client /usr/local/bin/psql /usr/local/bin/psql
+COPY --from=pg-client /usr/local/lib/libpq.so.5* /usr/local/lib/
+
 # Create non-root user
 RUN addgroup -g 1000 sub2api && \
     adduser -u 1000 -G sub2api -s /bin/sh -D sub2api
@@ -102,8 +121,9 @@ COPY --from=backend-builder --chown=sub2api:sub2api /app/backend/resources /app/
 # Create data directory
 RUN mkdir -p /app/data && chown sub2api:sub2api /app/data
 
-# Switch to non-root user
-USER sub2api
+# Copy entrypoint script (fixes volume permissions then drops to sub2api)
+COPY deploy/docker-entrypoint.sh /app/docker-entrypoint.sh
+RUN chmod +x /app/docker-entrypoint.sh
 
 # Expose port (can be overridden by SERVER_PORT env var)
 EXPOSE 8080
@@ -112,5 +132,6 @@ EXPOSE 8080
 HEALTHCHECK --interval=30s --timeout=10s --start-period=10s --retries=3 \
     CMD wget -q -T 5 -O /dev/null http://localhost:${SERVER_PORT:-8080}/health || exit 1
 
-# Run the application
-ENTRYPOINT ["/app/sub2api"]
+# Run the application (entrypoint fixes /app/data ownership then execs as sub2api)
+ENTRYPOINT ["/app/docker-entrypoint.sh"]
+CMD ["/app/sub2api"]

Dockerfile.goreleaser

@@ -5,7 +5,12 @@
 # It only packages the pre-built binary, no compilation needed.
 # =============================================================================
 
-FROM alpine:3.19
+ARG ALPINE_IMAGE=alpine:3.21
+ARG POSTGRES_IMAGE=postgres:18-alpine
+
+FROM ${POSTGRES_IMAGE} AS pg-client
+
+FROM ${ALPINE_IMAGE}
 
 LABEL maintainer="Wei-Shaw <github.com/Wei-Shaw>"
 LABEL description="Sub2API - AI API Gateway Platform"
@@ -16,8 +21,21 @@ RUN apk add --no-cache \
     ca-certificates \
     tzdata \
     curl \
+    su-exec \
+    libpq \
+    zstd-libs \
+    lz4-libs \
+    krb5-libs \
+    libldap \
+    libedit \
     && rm -rf /var/cache/apk/*
 
+# Copy pg_dump and psql from a version-matched PostgreSQL image so backup and
+# restore work in the runtime container without requiring Docker socket access.
+COPY --from=pg-client /usr/local/bin/pg_dump /usr/local/bin/pg_dump
+COPY --from=pg-client /usr/local/bin/psql /usr/local/bin/psql
+COPY --from=pg-client /usr/local/lib/libpq.so.5* /usr/local/lib/
+
 # Create non-root user
 RUN addgroup -g 1000 sub2api && \
     adduser -u 1000 -G sub2api -s /bin/sh -D sub2api
@@ -30,11 +48,15 @@ COPY sub2api /app/sub2api
 # Create data directory
 RUN mkdir -p /app/data && chown -R sub2api:sub2api /app
 
-USER sub2api
+# Copy entrypoint script (fixes volume permissions then drops to sub2api)
+COPY deploy/docker-entrypoint.sh /app/docker-entrypoint.sh
+RUN chmod +x /app/docker-entrypoint.sh
 
 EXPOSE 8080
 
 HEALTHCHECK --interval=30s --timeout=10s --start-period=10s --retries=3 \
     CMD curl -f http://localhost:${SERVER_PORT:-8080}/health || exit 1
 
-ENTRYPOINT ["/app/sub2api"]
+# Run the application (entrypoint fixes /app/data ownership then execs as sub2api)
+ENTRYPOINT ["/app/docker-entrypoint.sh"]
+CMD ["/app/sub2api"]

README.md

@@ -8,27 +8,31 @@
 [![Redis](https://img.shields.io/badge/Redis-7+-DC382D.svg)](https://redis.io/)
 [![Docker](https://img.shields.io/badge/Docker-Ready-2496ED.svg)](https://www.docker.com/)
 
+<a href="https://trendshift.io/repositories/21823" target="_blank"><img src="https://trendshift.io/api/badge/repositories/21823" alt="Wei-Shaw%2Fsub2api | Trendshift" width="250" height="55"/></a>
+
 **AI API Gateway Platform for Subscription Quota Distribution**
 
 English | [中文](README_CN.md)
 
 </div>
 
+> **Sub2API officially uses only the domains `sub2api.org` and `pincc.ai`. Other websites using the Sub2API name may be third-party deployments or services and are not affiliated with this project. Please verify and exercise your own judgment.**
+
 ---
 
 ## Demo
 
-Try Sub2API online: **https://demo.sub2api.org/**
+Try Sub2API online: **[https://demo.sub2api.org/](https://demo.sub2api.org/)**
 
 Demo credentials (shared demo environment; **not** created automatically for self-hosted installs):
 
 | Email | Password |
 |-------|----------|
-| admin@sub2api.com | admin123 |
+| admin@sub2api.org | admin123 |
 
 ## Overview
 
-Sub2API is an AI API gateway platform designed to distribute and manage API quotas from AI product subscriptions (like Claude Code $200/month). Users can access upstream AI services through platform-generated API Keys, while the platform handles authentication, billing, load balancing, and request forwarding.
+Sub2API is an AI API gateway platform designed to distribute and manage API quotas from AI product subscriptions. Users can access upstream AI services through platform-generated API Keys, while the platform handles authentication, billing, load balancing, and request forwarding.
 
 ## Features
 
@@ -39,6 +43,25 @@ Sub2API is an AI API gateway platform designed to distribute and manage API quot
 - **Concurrency Control** - Per-user and per-account concurrency limits
 - **Rate Limiting** - Configurable request and token rate limits
 - **Admin Dashboard** - Web interface for monitoring and management
+- **External System Integration** - Embed external systems (e.g. payment, ticketing) via iframe to extend the admin dashboard
+
+## Don't Want to Self-Host?
+
+<table>
+<tr>
+<td width="180" align="center" valign="middle"><a href="https://shop.pincc.ai/"><img src="assets/partners/logos/pincc-logo.png" alt="pincc" width="120"></a></td>
+<td valign="middle"><b><a href="https://shop.pincc.ai/">PinCC</a></b> is the official relay service built on Sub2API, offering stable access to Claude Code, Codex, Gemini and other popular models — ready to use, no deployment or maintenance required.</td>
+</tr>
+</table>
+
+## Ecosystem
+
+Community projects that extend or integrate with Sub2API:
+
+| Project | Description | Features |
+|---------|-------------|----------|
+| [Sub2ApiPay](https://github.com/touwaeriol/sub2apipay) | Self-service payment system | Self-service top-up and subscription purchase; supports YiPay protocol, WeChat Pay, Alipay, Stripe; embeddable via iframe |
+| [sub2api-mobile](https://github.com/ckken/sub2api-mobile) | Mobile admin console | Cross-platform app (iOS/Android/Web) for user management, account management, monitoring dashboard, and multi-backend switching; built with Expo + React Native |
 
 ## Tech Stack
 
@@ -51,10 +74,15 @@ Sub2API is an AI API gateway platform designed to distribute and manage API quot
 
 ---
 
-## Documentation
+## Nginx Reverse Proxy Note
 
-- Dependency Security: `docs/dependency-security.md`
-- Admin Payment Integration API: `docs/ADMIN_PAYMENT_INTEGRATION_API.md`
+When using Nginx as a reverse proxy for Sub2API (or CRS) with Codex CLI, add the following to the `http` block in your Nginx configuration:
+
+```nginx
+underscores_in_headers on;
+```
+
+Nginx drops headers containing underscores by default (e.g. `session_id`), which breaks sticky session routing in multi-account setups.
 
 ---
 

README_CN.md

@@ -8,27 +8,30 @@
 [![Redis](https://img.shields.io/badge/Redis-7+-DC382D.svg)](https://redis.io/)
 [![Docker](https://img.shields.io/badge/Docker-Ready-2496ED.svg)](https://www.docker.com/)
 
+<a href="https://trendshift.io/repositories/21823" target="_blank"><img src="https://trendshift.io/api/badge/repositories/21823" alt="Wei-Shaw%2Fsub2api | Trendshift" width="250" height="55"/></a>
+
 **AI API 网关平台 - 订阅配额分发管理**
 
 [English](README.md) | 中文
 
 </div>
 
+> **Sub2API 官方仅使用  `sub2api.org` 与 `pincc.ai` 两个域名。其他使用 Sub2API 名义的网站可能为第三方部署或服务，与本项目无关，请自行甄别。**
 ---
 
 ## 在线体验
 
-体验地址：**https://v2.pincc.ai/**
+体验地址：**[https://demo.sub2api.org/](https://demo.sub2api.org/)**
 
 演示账号（共享演示环境；自建部署不会自动创建该账号）：
 
 | 邮箱 | 密码 |
 |------|------|
-| admin@sub2api.com | admin123 |
+| admin@sub2api.org | admin123 |
 
 ## 项目概述
 
-Sub2API 是一个 AI API 网关平台，用于分发和管理 AI 产品订阅（如 Claude Code $200/月）的 API 配额。用户通过平台生成的 API Key 调用上游 AI 服务，平台负责鉴权、计费、负载均衡和请求转发。
+Sub2API 是一个 AI API 网关平台，用于分发和管理 AI 产品订阅的 API 配额。用户通过平台生成的 API Key 调用上游 AI 服务，平台负责鉴权、计费、负载均衡和请求转发。
 
 ## 核心功能
 
@@ -39,6 +42,25 @@ Sub2API 是一个 AI API 网关平台，用于分发和管理 AI 产品订阅（
 - **并发控制** - 用户级和账号级并发限制
 - **速率限制** - 可配置的请求和 Token 速率限制
 - **管理后台** - Web 界面进行监控和管理
+- **外部系统集成** - 支持通过 iframe 嵌入外部系统（如支付、工单等），扩展管理后台功能
+
+## 不想自建？试试官方中转
+
+<table>
+<tr>
+<td width="180" align="center" valign="middle"><a href="https://shop.pincc.ai/"><img src="assets/partners/logos/pincc-logo.png" alt="pincc" width="120"></a></td>
+<td valign="middle"><b><a href="https://shop.pincc.ai/">PinCC</a></b> 是基于 Sub2API 搭建的官方中转服务，提供 Claude Code、Codex、Gemini 等主流模型的稳定中转，开箱即用，免去自建部署与运维烦恼。</td>
+</tr>
+</table>
+
+## 生态项目
+
+围绕 Sub2API 的社区扩展与集成项目：
+
+| 项目 | 说明 | 功能 |
+|------|------|------|
+| [Sub2ApiPay](https://github.com/touwaeriol/sub2apipay) | 自助支付系统 | 用户自助充值、自助订阅购买；兼容易支付协议、微信官方支付、支付宝官方支付、Stripe；支持 iframe 嵌入管理后台 |
+| [sub2api-mobile](https://github.com/ckken/sub2api-mobile) | 移动端管理控制台 | 跨平台应用（iOS/Android/Web），支持用户管理、账号管理、监控看板、多后端切换；基于 Expo + React Native 构建 |
 
 ## 技术栈
 
@@ -51,17 +73,18 @@ Sub2API 是一个 AI API 网关平台，用于分发和管理 AI 产品订阅（
 
 ---
 
-## 文档
+## Nginx 反向代理注意事项
 
-- 依赖安全：`docs/dependency-security.md`
+通过 Nginx 反向代理 Sub2API（或 CRS 服务）并搭配 Codex CLI 使用时，需要在 Nginx 配置的 `http` 块中添加：
+
+```nginx
+underscores_in_headers on;
+```
+
+Nginx 默认会丢弃名称中含下划线的请求头（如 `session_id`），这会导致多账号环境下的粘性会话功能失效。
 
 ---
 
-## OpenAI Responses 兼容注意事项
-
-- 当请求包含 `function_call_output` 时，需要携带 `previous_response_id`，或在 `input` 中包含带 `call_id` 的 `tool_call`/`function_call`，或带非空 `id` 且与 `function_call_output.call_id` 匹配的 `item_reference`。
-- 若依赖上游历史记录，网关会强制 `store=true` 并需要复用 `previous_response_id`，以避免出现 “No tool call found for function call output” 错误。
-
 ## 部署方式
 
 ### 方式一：脚本安装（推荐）

backend/cmd/server/wire.go

@@ -41,6 +41,9 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 		// Server layer ProviderSet
 		server.ProviderSet,
 
+		// Privacy client factory for OpenAI training opt-out
+		providePrivacyClientFactory,
+
 		// BuildInfo provider
 		provideServiceBuildInfo,
 
@@ -53,6 +56,10 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	return nil, nil
 }
 
+func providePrivacyClientFactory() service.PrivacyClientFactory {
+	return repository.CreatePrivacyReqClient
+}
+
 func provideServiceBuildInfo(buildInfo handler.BuildInfo) service.BuildInfo {
 	return service.BuildInfo{
 		Version:   buildInfo.Version,
@@ -87,6 +94,7 @@ func provideCleanup(
 	antigravityOAuth *service.AntigravityOAuthService,
 	openAIGateway *service.OpenAIGatewayService,
 	scheduledTestRunner *service.ScheduledTestRunnerService,
+	backupSvc *service.BackupService,
 ) func() {
 	return func() {
 		ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
@@ -223,6 +231,12 @@ func provideCleanup(
 				}
 				return nil
 			}},
+			{"BackupService", func() error {
+				if backupSvc != nil {
+					backupSvc.Stop()
+				}
+				return nil
+			}},
 		}
 
 		infraSteps := []cleanupStep{

backend/cmd/server/wire_gen.go

@@ -81,6 +81,7 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	userHandler := handler.NewUserHandler(userService)
 	apiKeyHandler := handler.NewAPIKeyHandler(apiKeyService)
 	usageLogRepository := repository.NewUsageLogRepository(client, db)
+	usageBillingRepository := repository.NewUsageBillingRepository(client, db)
 	usageService := service.NewUsageService(usageLogRepository, userRepository, client, apiKeyAuthCacheInvalidator)
 	usageHandler := handler.NewUsageHandler(usageService, apiKeyService)
 	redeemHandler := handler.NewRedeemHandler(redeemService)
@@ -104,11 +105,11 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	proxyRepository := repository.NewProxyRepository(client, db)
 	proxyExitInfoProber := repository.NewProxyExitInfoProber(configConfig)
 	proxyLatencyCache := repository.NewProxyLatencyCache(redisClient)
-	adminService := service.NewAdminService(userRepository, groupRepository, accountRepository, soraAccountRepository, proxyRepository, apiKeyRepository, redeemCodeRepository, userGroupRateRepository, billingCacheService, proxyExitInfoProber, proxyLatencyCache, apiKeyAuthCacheInvalidator, client, settingService, subscriptionService, userSubscriptionRepository)
+	privacyClientFactory := providePrivacyClientFactory()
+	adminService := service.NewAdminService(userRepository, groupRepository, accountRepository, soraAccountRepository, proxyRepository, apiKeyRepository, redeemCodeRepository, userGroupRateRepository, billingCacheService, proxyExitInfoProber, proxyLatencyCache, apiKeyAuthCacheInvalidator, client, settingService, subscriptionService, userSubscriptionRepository, privacyClientFactory)
 	concurrencyCache := repository.ProvideConcurrencyCache(redisClient, configConfig)
 	concurrencyService := service.ProvideConcurrencyService(concurrencyCache, accountRepository, configConfig)
 	adminUserHandler := admin.NewUserHandler(adminService, concurrencyService)
-	groupHandler := admin.NewGroupHandler(adminService)
 	claudeOAuthClient := repository.NewClaudeOAuthClient()
 	oAuthService := service.NewOAuthService(proxyRepository, claudeOAuthClient)
 	openAIOAuthClient := repository.NewOpenAIOAuthClient()
@@ -122,6 +123,7 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	tempUnschedCache := repository.NewTempUnschedCache(redisClient)
 	timeoutCounterCache := repository.NewTimeoutCounterCache(redisClient)
 	geminiTokenCache := repository.NewGeminiTokenCache(redisClient)
+	oauthRefreshAPI := service.NewOAuthRefreshAPI(accountRepository, geminiTokenCache)
 	compositeTokenCacheInvalidator := service.NewCompositeTokenCacheInvalidator(geminiTokenCache)
 	rateLimitService := service.ProvideRateLimitService(accountRepository, usageLogRepository, configConfig, geminiQuotaService, tempUnschedCache, timeoutCounterCache, settingService, compositeTokenCacheInvalidator)
 	httpUpstream := repository.NewHTTPUpstream(configConfig)
@@ -130,20 +132,26 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	usageCache := service.NewUsageCache()
 	identityCache := repository.NewIdentityCache(redisClient)
 	accountUsageService := service.NewAccountUsageService(accountRepository, usageLogRepository, claudeUsageFetcher, geminiQuotaService, antigravityQuotaFetcher, usageCache, identityCache)
-	geminiTokenProvider := service.NewGeminiTokenProvider(accountRepository, geminiTokenCache, geminiOAuthService)
+	geminiTokenProvider := service.ProvideGeminiTokenProvider(accountRepository, geminiTokenCache, geminiOAuthService, oauthRefreshAPI)
 	gatewayCache := repository.NewGatewayCache(redisClient)
 	schedulerOutboxRepository := repository.NewSchedulerOutboxRepository(db)
 	schedulerSnapshotService := service.ProvideSchedulerSnapshotService(schedulerCache, schedulerOutboxRepository, accountRepository, groupRepository, configConfig)
-	antigravityTokenProvider := service.NewAntigravityTokenProvider(accountRepository, geminiTokenCache, antigravityOAuthService)
+	antigravityTokenProvider := service.ProvideAntigravityTokenProvider(accountRepository, geminiTokenCache, antigravityOAuthService, oauthRefreshAPI)
 	antigravityGatewayService := service.NewAntigravityGatewayService(accountRepository, gatewayCache, schedulerSnapshotService, antigravityTokenProvider, rateLimitService, httpUpstream, settingService)
 	accountTestService := service.NewAccountTestService(accountRepository, geminiTokenProvider, antigravityGatewayService, httpUpstream, configConfig)
 	crsSyncService := service.NewCRSSyncService(accountRepository, proxyRepository, oAuthService, openAIOAuthService, geminiOAuthService, configConfig)
 	sessionLimitCache := repository.ProvideSessionLimitCache(redisClient, configConfig)
 	rpmCache := repository.NewRPMCache(redisClient)
+	groupCapacityService := service.NewGroupCapacityService(accountRepository, groupRepository, concurrencyService, sessionLimitCache, rpmCache)
+	groupHandler := admin.NewGroupHandler(adminService, dashboardService, groupCapacityService)
 	accountHandler := admin.NewAccountHandler(adminService, oAuthService, openAIOAuthService, geminiOAuthService, antigravityOAuthService, rateLimitService, accountUsageService, accountTestService, concurrencyService, crsSyncService, sessionLimitCache, rpmCache, compositeTokenCacheInvalidator)
 	adminAnnouncementHandler := admin.NewAnnouncementHandler(announcementService)
 	dataManagementService := service.NewDataManagementService()
 	dataManagementHandler := admin.NewDataManagementHandler(dataManagementService)
+	backupObjectStoreFactory := repository.NewS3BackupStoreFactory()
+	dbDumper := repository.NewPgDumper(configConfig)
+	backupService := service.ProvideBackupService(settingRepository, configConfig, secretEncryptor, backupObjectStoreFactory, dbDumper)
+	backupHandler := admin.NewBackupHandler(backupService, userService)
 	oAuthHandler := admin.NewOAuthHandler(oAuthService)
 	openAIOAuthHandler := admin.NewOpenAIOAuthHandler(openAIOAuthService, adminService)
 	geminiOAuthHandler := admin.NewGeminiOAuthHandler(geminiOAuthService)
@@ -160,11 +168,11 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	billingService := service.NewBillingService(configConfig, pricingService)
 	identityService := service.NewIdentityService(identityCache)
 	deferredService := service.ProvideDeferredService(accountRepository, timingWheelService)
-	claudeTokenProvider := service.NewClaudeTokenProvider(accountRepository, geminiTokenCache, oAuthService)
+	claudeTokenProvider := service.ProvideClaudeTokenProvider(accountRepository, geminiTokenCache, oAuthService, oauthRefreshAPI)
 	digestSessionStore := service.NewDigestSessionStore()
-	gatewayService := service.NewGatewayService(accountRepository, groupRepository, usageLogRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, identityService, httpUpstream, deferredService, claudeTokenProvider, sessionLimitCache, rpmCache, digestSessionStore, settingService)
-	openAITokenProvider := service.NewOpenAITokenProvider(accountRepository, geminiTokenCache, openAIOAuthService)
-	openAIGatewayService := service.NewOpenAIGatewayService(accountRepository, usageLogRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, httpUpstream, deferredService, openAITokenProvider)
+	gatewayService := service.NewGatewayService(accountRepository, groupRepository, usageLogRepository, usageBillingRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, identityService, httpUpstream, deferredService, claudeTokenProvider, sessionLimitCache, rpmCache, digestSessionStore, settingService)
+	openAITokenProvider := service.ProvideOpenAITokenProvider(accountRepository, geminiTokenCache, openAIOAuthService, oauthRefreshAPI)
+	openAIGatewayService := service.NewOpenAIGatewayService(accountRepository, usageLogRepository, usageBillingRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, httpUpstream, deferredService, openAITokenProvider)
 	geminiMessagesCompatService := service.NewGeminiMessagesCompatService(accountRepository, groupRepository, gatewayCache, schedulerSnapshotService, geminiTokenProvider, rateLimitService, httpUpstream, antigravityGatewayService, configConfig)
 	opsSystemLogSink := service.ProvideOpsSystemLogSink(opsRepository)
 	opsService := service.NewOpsService(opsRepository, settingRepository, configConfig, accountRepository, userRepository, concurrencyService, gatewayService, openAIGatewayService, geminiMessagesCompatService, antigravityGatewayService, opsSystemLogSink)
@@ -199,7 +207,7 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	scheduledTestResultRepository := repository.NewScheduledTestResultRepository(db)
 	scheduledTestService := service.ProvideScheduledTestService(scheduledTestPlanRepository, scheduledTestResultRepository)
 	scheduledTestHandler := admin.NewScheduledTestHandler(scheduledTestService)
-	adminHandlers := handler.ProvideAdminHandlers(dashboardHandler, adminUserHandler, groupHandler, accountHandler, adminAnnouncementHandler, dataManagementHandler, oAuthHandler, openAIOAuthHandler, geminiOAuthHandler, antigravityOAuthHandler, proxyHandler, adminRedeemHandler, promoHandler, settingHandler, opsHandler, systemHandler, adminSubscriptionHandler, adminUsageHandler, userAttributeHandler, errorPassthroughHandler, adminAPIKeyHandler, scheduledTestHandler)
+	adminHandlers := handler.ProvideAdminHandlers(dashboardHandler, adminUserHandler, groupHandler, accountHandler, adminAnnouncementHandler, dataManagementHandler, backupHandler, oAuthHandler, openAIOAuthHandler, geminiOAuthHandler, antigravityOAuthHandler, proxyHandler, adminRedeemHandler, promoHandler, settingHandler, opsHandler, systemHandler, adminSubscriptionHandler, adminUsageHandler, userAttributeHandler, errorPassthroughHandler, adminAPIKeyHandler, scheduledTestHandler)
 	usageRecordWorkerPool := service.NewUsageRecordWorkerPool(configConfig)
 	userMsgQueueCache := repository.NewUserMsgQueueCache(redisClient)
 	userMessageQueueService := service.ProvideUserMessageQueueService(userMsgQueueCache, rpmCache, configConfig)
@@ -226,11 +234,11 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
 	opsCleanupService := service.ProvideOpsCleanupService(opsRepository, db, redisClient, configConfig)
 	opsScheduledReportService := service.ProvideOpsScheduledReportService(opsService, userService, emailService, redisClient, configConfig)
 	soraMediaCleanupService := service.ProvideSoraMediaCleanupService(soraMediaStorage, configConfig)
-	tokenRefreshService := service.ProvideTokenRefreshService(accountRepository, soraAccountRepository, oAuthService, openAIOAuthService, geminiOAuthService, antigravityOAuthService, compositeTokenCacheInvalidator, schedulerCache, configConfig, tempUnschedCache)
+	tokenRefreshService := service.ProvideTokenRefreshService(accountRepository, soraAccountRepository, oAuthService, openAIOAuthService, geminiOAuthService, antigravityOAuthService, compositeTokenCacheInvalidator, schedulerCache, configConfig, tempUnschedCache, privacyClientFactory, proxyRepository, oauthRefreshAPI)
 	accountExpiryService := service.ProvideAccountExpiryService(accountRepository)
 	subscriptionExpiryService := service.ProvideSubscriptionExpiryService(userSubscriptionRepository)
 	scheduledTestRunnerService := service.ProvideScheduledTestRunnerService(scheduledTestPlanRepository, scheduledTestService, accountTestService, rateLimitService, configConfig)
-	v := provideCleanup(client, redisClient, opsMetricsCollector, opsAggregationService, opsAlertEvaluatorService, opsCleanupService, opsScheduledReportService, opsSystemLogSink, soraMediaCleanupService, schedulerSnapshotService, tokenRefreshService, accountExpiryService, subscriptionExpiryService, usageCleanupService, idempotencyCleanupService, pricingService, emailQueueService, billingCacheService, usageRecordWorkerPool, subscriptionService, oAuthService, openAIOAuthService, geminiOAuthService, antigravityOAuthService, openAIGatewayService, scheduledTestRunnerService)
+	v := provideCleanup(client, redisClient, opsMetricsCollector, opsAggregationService, opsAlertEvaluatorService, opsCleanupService, opsScheduledReportService, opsSystemLogSink, soraMediaCleanupService, schedulerSnapshotService, tokenRefreshService, accountExpiryService, subscriptionExpiryService, usageCleanupService, idempotencyCleanupService, pricingService, emailQueueService, billingCacheService, usageRecordWorkerPool, subscriptionService, oAuthService, openAIOAuthService, geminiOAuthService, antigravityOAuthService, openAIGatewayService, scheduledTestRunnerService, backupService)
 	application := &Application{
 		Server:  httpServer,
 		Cleanup: v,
@@ -245,6 +253,10 @@ type Application struct {
 	Cleanup func()
 }
 
+func providePrivacyClientFactory() service.PrivacyClientFactory {
+	return repository.CreatePrivacyReqClient
+}
+
 func provideServiceBuildInfo(buildInfo handler.BuildInfo) service.BuildInfo {
 	return service.BuildInfo{
 		Version:   buildInfo.Version,
@@ -279,6 +291,7 @@ func provideCleanup(
 	antigravityOAuth *service.AntigravityOAuthService,
 	openAIGateway *service.OpenAIGatewayService,
 	scheduledTestRunner *service.ScheduledTestRunnerService,
+	backupSvc *service.BackupService,
 ) func() {
 	return func() {
 		ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
@@ -414,6 +427,12 @@ func provideCleanup(
 				}
 				return nil
 			}},
+			{"BackupService", func() error {
+				if backupSvc != nil {
+					backupSvc.Stop()
+				}
+				return nil
+			}},
 		}
 
 		infraSteps := []cleanupStep{

backend/cmd/server/wire_gen_test.go

@@ -75,6 +75,7 @@ func TestProvideCleanup_WithMinimalDependencies_NoPanic(t *testing.T) {
 		antigravityOAuthSvc,
 		nil, // openAIGateway
 		nil, // scheduledTestRunner
+		nil, // backupSvc
 	)
 
 	require.NotPanics(t, func() {

backend/go.mod

@@ -7,7 +7,7 @@ require (
 	github.com/DATA-DOG/go-sqlmock v1.5.2
 	github.com/DouDOU-start/go-sora2api v1.1.0
 	github.com/alitto/pond/v2 v2.6.2
-	github.com/aws/aws-sdk-go-v2 v1.41.2
+	github.com/aws/aws-sdk-go-v2 v1.41.3
 	github.com/aws/aws-sdk-go-v2/config v1.32.10
 	github.com/aws/aws-sdk-go-v2/credentials v1.19.10
 	github.com/aws/aws-sdk-go-v2/service/s3 v1.96.2
@@ -66,7 +66,7 @@ require (
 	github.com/aws/aws-sdk-go-v2/service/sso v1.30.11 // indirect
 	github.com/aws/aws-sdk-go-v2/service/ssooidc v1.35.15 // indirect
 	github.com/aws/aws-sdk-go-v2/service/sts v1.41.7 // indirect
-	github.com/aws/smithy-go v1.24.1 // indirect
+	github.com/aws/smithy-go v1.24.2 // indirect
 	github.com/bdandy/go-errors v1.2.2 // indirect
 	github.com/bdandy/go-socks4 v1.2.3 // indirect
 	github.com/bmatcuk/doublestar v1.3.4 // indirect

backend/go.sum

@@ -24,6 +24,8 @@ github.com/apparentlymart/go-textseg/v15 v15.0.0 h1:uYvfpb3DyLSCGWnctWKGj857c6ew
 github.com/apparentlymart/go-textseg/v15 v15.0.0/go.mod h1:K8XmNZdhEBkdlyDdvbmmsvpAG721bKi0joRfFdHIWJ4=
 github.com/aws/aws-sdk-go-v2 v1.41.2 h1:LuT2rzqNQsauaGkPK/7813XxcZ3o3yePY0Iy891T2ls=
 github.com/aws/aws-sdk-go-v2 v1.41.2/go.mod h1:IvvlAZQXvTXznUPfRVfryiG1fbzE2NGK6m9u39YQ+S4=
+github.com/aws/aws-sdk-go-v2 v1.41.3 h1:4kQ/fa22KjDt13QCy1+bYADvdgcxpfH18f0zP542kZA=
+github.com/aws/aws-sdk-go-v2 v1.41.3/go.mod h1:mwsPRE8ceUUpiTgF7QmQIJ7lgsKUPQOUl3o72QBrE1o=
 github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.5 h1:zWFmPmgw4sveAYi1mRqG+E/g0461cJ5M4bJ8/nc6d3Q=
 github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.7.5/go.mod h1:nVUlMLVV8ycXSb7mSkcNu9e3v/1TJq2RTlrPwhYWr5c=
 github.com/aws/aws-sdk-go-v2/config v1.32.10 h1:9DMthfO6XWZYLfzZglAgW5Fyou2nRI5CuV44sTedKBI=
@@ -60,6 +62,8 @@ github.com/aws/aws-sdk-go-v2/service/sts v1.41.7 h1:NITQpgo9A5NrDZ57uOWj+abvXSb8
 github.com/aws/aws-sdk-go-v2/service/sts v1.41.7/go.mod h1:sks5UWBhEuWYDPdwlnRFn1w7xWdH29Jcpe+/PJQefEs=
 github.com/aws/smithy-go v1.24.1 h1:VbyeNfmYkWoxMVpGUAbQumkODcYmfMRfZ8yQiH30SK0=
 github.com/aws/smithy-go v1.24.1/go.mod h1:LEj2LM3rBRQJxPZTB4KuzZkaZYnZPnvgIhb4pu07mx0=
+github.com/aws/smithy-go v1.24.2 h1:FzA3bu/nt/vDvmnkg+R8Xl46gmzEDam6mZ1hzmwXFng=
+github.com/aws/smithy-go v1.24.2/go.mod h1:YE2RhdIuDbA5E5bTdciG9KrW3+TiEONeUWCqxX9i1Fc=
 github.com/bdandy/go-errors v1.2.2 h1:WdFv/oukjTJCLa79UfkGmwX7ZxONAihKu4V0mLIs11Q=
 github.com/bdandy/go-errors v1.2.2/go.mod h1:NkYHl4Fey9oRRdbB1CoC6e84tuqQHiqrOcZpqFEkBxM=
 github.com/bdandy/go-socks4 v1.2.3 h1:Q6Y2heY1GRjCtHbmlKfnwrKVU/k81LS8mRGLRlmDlic=

backend/internal/config/config.go

@@ -934,9 +934,10 @@ type DashboardAggregationConfig struct {
 
 // DashboardAggregationRetentionConfig 预聚合保留窗口
 type DashboardAggregationRetentionConfig struct {
-	UsageLogsDays int `mapstructure:"usage_logs_days"`
-	HourlyDays    int `mapstructure:"hourly_days"`
-	DailyDays     int `mapstructure:"daily_days"`
+	UsageLogsDays         int `mapstructure:"usage_logs_days"`
+	UsageBillingDedupDays int `mapstructure:"usage_billing_dedup_days"`
+	HourlyDays            int `mapstructure:"hourly_days"`
+	DailyDays             int `mapstructure:"daily_days"`
 }
 
 // UsageCleanupConfig 使用记录清理任务配置
@@ -1301,6 +1302,7 @@ func setDefaults() {
 	viper.SetDefault("dashboard_aggregation.backfill_enabled", false)
 	viper.SetDefault("dashboard_aggregation.backfill_max_days", 31)
 	viper.SetDefault("dashboard_aggregation.retention.usage_logs_days", 90)
+	viper.SetDefault("dashboard_aggregation.retention.usage_billing_dedup_days", 365)
 	viper.SetDefault("dashboard_aggregation.retention.hourly_days", 180)
 	viper.SetDefault("dashboard_aggregation.retention.daily_days", 730)
 	viper.SetDefault("dashboard_aggregation.recompute_days", 2)
@@ -1758,6 +1760,12 @@ func (c *Config) Validate() error {
 		if c.DashboardAgg.Retention.UsageLogsDays <= 0 {
 			return fmt.Errorf("dashboard_aggregation.retention.usage_logs_days must be positive")
 		}
+		if c.DashboardAgg.Retention.UsageBillingDedupDays <= 0 {
+			return fmt.Errorf("dashboard_aggregation.retention.usage_billing_dedup_days must be positive")
+		}
+		if c.DashboardAgg.Retention.UsageBillingDedupDays < c.DashboardAgg.Retention.UsageLogsDays {
+			return fmt.Errorf("dashboard_aggregation.retention.usage_billing_dedup_days must be greater than or equal to usage_logs_days")
+		}
 		if c.DashboardAgg.Retention.HourlyDays <= 0 {
 			return fmt.Errorf("dashboard_aggregation.retention.hourly_days must be positive")
 		}
@@ -1780,6 +1788,14 @@ func (c *Config) Validate() error {
 		if c.DashboardAgg.Retention.UsageLogsDays < 0 {
 			return fmt.Errorf("dashboard_aggregation.retention.usage_logs_days must be non-negative")
 		}
+		if c.DashboardAgg.Retention.UsageBillingDedupDays < 0 {
+			return fmt.Errorf("dashboard_aggregation.retention.usage_billing_dedup_days must be non-negative")
+		}
+		if c.DashboardAgg.Retention.UsageBillingDedupDays > 0 &&
+			c.DashboardAgg.Retention.UsageLogsDays > 0 &&
+			c.DashboardAgg.Retention.UsageBillingDedupDays < c.DashboardAgg.Retention.UsageLogsDays {
+			return fmt.Errorf("dashboard_aggregation.retention.usage_billing_dedup_days must be greater than or equal to usage_logs_days")
+		}
 		if c.DashboardAgg.Retention.HourlyDays < 0 {
 			return fmt.Errorf("dashboard_aggregation.retention.hourly_days must be non-negative")
 		}

backend/internal/config/config_test.go

@@ -441,6 +441,9 @@ func TestLoadDefaultDashboardAggregationConfig(t *testing.T) {
 	if cfg.DashboardAgg.Retention.UsageLogsDays != 90 {
 		t.Fatalf("DashboardAgg.Retention.UsageLogsDays = %d, want 90", cfg.DashboardAgg.Retention.UsageLogsDays)
 	}
+	if cfg.DashboardAgg.Retention.UsageBillingDedupDays != 365 {
+		t.Fatalf("DashboardAgg.Retention.UsageBillingDedupDays = %d, want 365", cfg.DashboardAgg.Retention.UsageBillingDedupDays)
+	}
 	if cfg.DashboardAgg.Retention.HourlyDays != 180 {
 		t.Fatalf("DashboardAgg.Retention.HourlyDays = %d, want 180", cfg.DashboardAgg.Retention.HourlyDays)
 	}
@@ -1016,6 +1019,23 @@ func TestValidateConfigErrors(t *testing.T) {
 			mutate:  func(c *Config) { c.DashboardAgg.Enabled = true; c.DashboardAgg.Retention.UsageLogsDays = 0 },
 			wantErr: "dashboard_aggregation.retention.usage_logs_days",
 		},
+		{
+			name: "dashboard aggregation dedup retention",
+			mutate: func(c *Config) {
+				c.DashboardAgg.Enabled = true
+				c.DashboardAgg.Retention.UsageBillingDedupDays = 0
+			},
+			wantErr: "dashboard_aggregation.retention.usage_billing_dedup_days",
+		},
+		{
+			name: "dashboard aggregation dedup retention smaller than usage logs",
+			mutate: func(c *Config) {
+				c.DashboardAgg.Enabled = true
+				c.DashboardAgg.Retention.UsageLogsDays = 30
+				c.DashboardAgg.Retention.UsageBillingDedupDays = 29
+			},
+			wantErr: "dashboard_aggregation.retention.usage_billing_dedup_days",
+		},
 		{
 			name:    "dashboard aggregation disabled interval",
 			mutate:  func(c *Config) { c.DashboardAgg.Enabled = false; c.DashboardAgg.IntervalSeconds = -1 },

backend/internal/domain/constants.go

@@ -31,6 +31,7 @@ const (
 	AccountTypeSetupToken = "setup-token" // Setup Token类型账号（inference only scope）
 	AccountTypeAPIKey     = "apikey"      // API Key类型账号
 	AccountTypeUpstream   = "upstream"    // 上游透传类型账号（通过 Base URL + API Key 连接上游）
+	AccountTypeBedrock    = "bedrock"     // AWS Bedrock 类型账号（通过 SigV4 签名或 API Key 连接 Bedrock，由 credentials.auth_mode 区分）
 )
 
 // Redeem type constants
@@ -113,3 +114,27 @@ var DefaultAntigravityModelMapping = map[string]string{
 	"gpt-oss-120b-medium":    "gpt-oss-120b-medium",
 	"tab_flash_lite_preview": "tab_flash_lite_preview",
 }
+
+// DefaultBedrockModelMapping 是 AWS Bedrock 平台的默认模型映射
+// 将 Anthropic 标准模型名映射到 Bedrock 模型 ID
+// 注意：此处的 "us." 前缀仅为默认值，ResolveBedrockModelID 会根据账号配置的
+// aws_region 自动调整为匹配的区域前缀（如 eu.、apac.、jp. 等）
+var DefaultBedrockModelMapping = map[string]string{
+	// Claude Opus
+	"claude-opus-4-6-thinking": "us.anthropic.claude-opus-4-6-v1",
+	"claude-opus-4-6":          "us.anthropic.claude-opus-4-6-v1",
+	"claude-opus-4-5-thinking": "us.anthropic.claude-opus-4-5-20251101-v1:0",
+	"claude-opus-4-5-20251101": "us.anthropic.claude-opus-4-5-20251101-v1:0",
+	"claude-opus-4-1":          "us.anthropic.claude-opus-4-1-20250805-v1:0",
+	"claude-opus-4-20250514":   "us.anthropic.claude-opus-4-20250514-v1:0",
+	// Claude Sonnet
+	"claude-sonnet-4-6-thinking": "us.anthropic.claude-sonnet-4-6",
+	"claude-sonnet-4-6":          "us.anthropic.claude-sonnet-4-6",
+	"claude-sonnet-4-5":          "us.anthropic.claude-sonnet-4-5-20250929-v1:0",
+	"claude-sonnet-4-5-thinking": "us.anthropic.claude-sonnet-4-5-20250929-v1:0",
+	"claude-sonnet-4-5-20250929": "us.anthropic.claude-sonnet-4-5-20250929-v1:0",
+	"claude-sonnet-4-20250514":   "us.anthropic.claude-sonnet-4-20250514-v1:0",
+	// Claude Haiku
+	"claude-haiku-4-5":          "us.anthropic.claude-haiku-4-5-20251001-v1:0",
+	"claude-haiku-4-5-20251001": "us.anthropic.claude-haiku-4-5-20251001-v1:0",
+}

backend/internal/handler/admin/account_handler.go

@@ -97,7 +97,7 @@ type CreateAccountRequest struct {
 	Name                    string         `json:"name" binding:"required"`
 	Notes                   *string        `json:"notes"`
 	Platform                string         `json:"platform" binding:"required"`
-	Type                    string         `json:"type" binding:"required,oneof=oauth setup-token apikey upstream"`
+	Type                    string         `json:"type" binding:"required,oneof=oauth setup-token apikey upstream bedrock"`
 	Credentials             map[string]any `json:"credentials" binding:"required"`
 	Extra                   map[string]any `json:"extra"`
 	ProxyID                 *int64         `json:"proxy_id"`
@@ -116,7 +116,7 @@ type CreateAccountRequest struct {
 type UpdateAccountRequest struct {
 	Name                    string         `json:"name"`
 	Notes                   *string        `json:"notes"`
-	Type                    string         `json:"type" binding:"omitempty,oneof=oauth setup-token apikey upstream"`
+	Type                    string         `json:"type" binding:"omitempty,oneof=oauth setup-token apikey upstream bedrock"`
 	Credentials             map[string]any `json:"credentials"`
 	Extra                   map[string]any `json:"extra"`
 	ProxyID                 *int64         `json:"proxy_id"`
@@ -865,6 +865,9 @@ func (h *AccountHandler) refreshSingleAccount(ctx context.Context, account *serv
 		}
 	}
 
+	// OpenAI OAuth: 刷新成功后检查并设置 privacy_mode
+	h.adminService.EnsureOpenAIPrivacy(ctx, updatedAccount)
+
 	return updatedAccount, "", nil
 }
 
@@ -1715,13 +1718,12 @@ func (h *AccountHandler) GetAvailableModels(c *gin.Context) {
 
 	// Handle OpenAI accounts
 	if account.IsOpenAI() {
-		// For OAuth accounts: return default OpenAI models
-		if account.IsOAuth() {
+		// OpenAI 自动透传会绕过常规模型改写，测试/模型列表也应回落到默认模型集。
+		if account.IsOpenAIPassthroughEnabled() {
 			response.Success(c, openai.DefaultModels)
 			return
 		}
 
-		// For API Key accounts: check model_mapping
 		mapping := account.GetModelMapping()
 		if len(mapping) == 0 {
 			response.Success(c, openai.DefaultModels)

backend/internal/handler/admin/account_handler_available_models_test.go

@@ -0,0 +1,105 @@
+package admin
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/require"
+)
+
+type availableModelsAdminService struct {
+	*stubAdminService
+	account service.Account
+}
+
+func (s *availableModelsAdminService) GetAccount(_ context.Context, id int64) (*service.Account, error) {
+	if s.account.ID == id {
+		acc := s.account
+		return &acc, nil
+	}
+	return s.stubAdminService.GetAccount(context.Background(), id)
+}
+
+func setupAvailableModelsRouter(adminSvc service.AdminService) *gin.Engine {
+	gin.SetMode(gin.TestMode)
+	router := gin.New()
+	handler := NewAccountHandler(adminSvc, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil)
+	router.GET("/api/v1/admin/accounts/:id/models", handler.GetAvailableModels)
+	return router
+}
+
+func TestAccountHandlerGetAvailableModels_OpenAIOAuthUsesExplicitModelMapping(t *testing.T) {
+	svc := &availableModelsAdminService{
+		stubAdminService: newStubAdminService(),
+		account: service.Account{
+			ID:       42,
+			Name:     "openai-oauth",
+			Platform: service.PlatformOpenAI,
+			Type:     service.AccountTypeOAuth,
+			Status:   service.StatusActive,
+			Credentials: map[string]any{
+				"model_mapping": map[string]any{
+					"gpt-5": "gpt-5.1",
+				},
+			},
+		},
+	}
+	router := setupAvailableModelsRouter(svc)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodGet, "/api/v1/admin/accounts/42/models", nil)
+	router.ServeHTTP(rec, req)
+
+	require.Equal(t, http.StatusOK, rec.Code)
+
+	var resp struct {
+		Data []struct {
+			ID string `json:"id"`
+		} `json:"data"`
+	}
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &resp))
+	require.Len(t, resp.Data, 1)
+	require.Equal(t, "gpt-5", resp.Data[0].ID)
+}
+
+func TestAccountHandlerGetAvailableModels_OpenAIOAuthPassthroughFallsBackToDefaults(t *testing.T) {
+	svc := &availableModelsAdminService{
+		stubAdminService: newStubAdminService(),
+		account: service.Account{
+			ID:       43,
+			Name:     "openai-oauth-passthrough",
+			Platform: service.PlatformOpenAI,
+			Type:     service.AccountTypeOAuth,
+			Status:   service.StatusActive,
+			Credentials: map[string]any{
+				"model_mapping": map[string]any{
+					"gpt-5": "gpt-5.1",
+				},
+			},
+			Extra: map[string]any{
+				"openai_passthrough": true,
+			},
+		},
+	}
+	router := setupAvailableModelsRouter(svc)
+
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodGet, "/api/v1/admin/accounts/43/models", nil)
+	router.ServeHTTP(rec, req)
+
+	require.Equal(t, http.StatusOK, rec.Code)
+
+	var resp struct {
+		Data []struct {
+			ID string `json:"id"`
+		} `json:"data"`
+	}
+	require.NoError(t, json.Unmarshal(rec.Body.Bytes(), &resp))
+	require.NotEmpty(t, resp.Data)
+	require.NotEqual(t, "gpt-5", resp.Data[0].ID)
+}

backend/internal/handler/admin/admin_basic_handlers_test.go

@@ -17,7 +17,7 @@ func setupAdminRouter() (*gin.Engine, *stubAdminService) {
 	adminSvc := newStubAdminService()
 
 	userHandler := NewUserHandler(adminSvc, nil)
-	groupHandler := NewGroupHandler(adminSvc)
+	groupHandler := NewGroupHandler(adminSvc, nil, nil)
 	proxyHandler := NewProxyHandler(adminSvc)
 	redeemHandler := NewRedeemHandler(adminSvc, nil)
 

backend/internal/handler/admin/admin_service_stub_test.go

@@ -175,6 +175,18 @@ func (s *stubAdminService) GetGroupAPIKeys(ctx context.Context, groupID int64, p
 	return s.apiKeys, int64(len(s.apiKeys)), nil
 }
 
+func (s *stubAdminService) GetGroupRateMultipliers(_ context.Context, _ int64) ([]service.UserGroupRateEntry, error) {
+	return nil, nil
+}
+
+func (s *stubAdminService) ClearGroupRateMultipliers(_ context.Context, _ int64) error {
+	return nil
+}
+
+func (s *stubAdminService) BatchSetGroupRateMultipliers(_ context.Context, _ int64, _ []service.GroupRateMultiplierInput) error {
+	return nil
+}
+
 func (s *stubAdminService) ListAccounts(ctx context.Context, page, pageSize int, platform, accountType, status, search string, groupID int64) ([]service.Account, int64, error) {
 	return s.accounts, int64(len(s.accounts)), nil
 }
@@ -429,5 +441,9 @@ func (s *stubAdminService) ResetAccountQuota(ctx context.Context, id int64) erro
 	return nil
 }
 
+func (s *stubAdminService) EnsureOpenAIPrivacy(ctx context.Context, account *service.Account) string {
+	return ""
+}
+
 // Ensure stub implements interface.
 var _ service.AdminService = (*stubAdminService)(nil)

backend/internal/handler/admin/backup_handler.go

@@ -0,0 +1,205 @@
+package admin
+
+import (
+	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
+	"github.com/Wei-Shaw/sub2api/internal/server/middleware"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+)
+
+type BackupHandler struct {
+	backupService *service.BackupService
+	userService   *service.UserService
+}
+
+func NewBackupHandler(backupService *service.BackupService, userService *service.UserService) *BackupHandler {
+	return &BackupHandler{
+		backupService: backupService,
+		userService:   userService,
+	}
+}
+
+// ─── S3 配置 ───
+
+func (h *BackupHandler) GetS3Config(c *gin.Context) {
+	cfg, err := h.backupService.GetS3Config(c.Request.Context())
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, cfg)
+}
+
+func (h *BackupHandler) UpdateS3Config(c *gin.Context) {
+	var req service.BackupS3Config
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, "Invalid request: "+err.Error())
+		return
+	}
+	cfg, err := h.backupService.UpdateS3Config(c.Request.Context(), req)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, cfg)
+}
+
+func (h *BackupHandler) TestS3Connection(c *gin.Context) {
+	var req service.BackupS3Config
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, "Invalid request: "+err.Error())
+		return
+	}
+	err := h.backupService.TestS3Connection(c.Request.Context(), req)
+	if err != nil {
+		response.Success(c, gin.H{"ok": false, "message": err.Error()})
+		return
+	}
+	response.Success(c, gin.H{"ok": true, "message": "connection successful"})
+}
+
+// ─── 定时备份 ───
+
+func (h *BackupHandler) GetSchedule(c *gin.Context) {
+	cfg, err := h.backupService.GetSchedule(c.Request.Context())
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, cfg)
+}
+
+func (h *BackupHandler) UpdateSchedule(c *gin.Context) {
+	var req service.BackupScheduleConfig
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, "Invalid request: "+err.Error())
+		return
+	}
+	cfg, err := h.backupService.UpdateSchedule(c.Request.Context(), req)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, cfg)
+}
+
+// ─── 备份操作 ───
+
+type CreateBackupRequest struct {
+	ExpireDays *int `json:"expire_days"` // nil=使用默认值14，0=永不过期
+}
+
+func (h *BackupHandler) CreateBackup(c *gin.Context) {
+	var req CreateBackupRequest
+	_ = c.ShouldBindJSON(&req) // 允许空 body
+
+	expireDays := 14 // 默认14天过期
+	if req.ExpireDays != nil {
+		expireDays = *req.ExpireDays
+	}
+
+	record, err := h.backupService.StartBackup(c.Request.Context(), "manual", expireDays)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Accepted(c, record)
+}
+
+func (h *BackupHandler) ListBackups(c *gin.Context) {
+	records, err := h.backupService.ListBackups(c.Request.Context())
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	if records == nil {
+		records = []service.BackupRecord{}
+	}
+	response.Success(c, gin.H{"items": records})
+}
+
+func (h *BackupHandler) GetBackup(c *gin.Context) {
+	backupID := c.Param("id")
+	if backupID == "" {
+		response.BadRequest(c, "backup ID is required")
+		return
+	}
+	record, err := h.backupService.GetBackupRecord(c.Request.Context(), backupID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, record)
+}
+
+func (h *BackupHandler) DeleteBackup(c *gin.Context) {
+	backupID := c.Param("id")
+	if backupID == "" {
+		response.BadRequest(c, "backup ID is required")
+		return
+	}
+	if err := h.backupService.DeleteBackup(c.Request.Context(), backupID); err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, gin.H{"deleted": true})
+}
+
+func (h *BackupHandler) GetDownloadURL(c *gin.Context) {
+	backupID := c.Param("id")
+	if backupID == "" {
+		response.BadRequest(c, "backup ID is required")
+		return
+	}
+	url, err := h.backupService.GetBackupDownloadURL(c.Request.Context(), backupID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Success(c, gin.H{"url": url})
+}
+
+// ─── 恢复操作（需要重新输入管理员密码） ───
+
+type RestoreBackupRequest struct {
+	Password string `json:"password" binding:"required"`
+}
+
+func (h *BackupHandler) RestoreBackup(c *gin.Context) {
+	backupID := c.Param("id")
+	if backupID == "" {
+		response.BadRequest(c, "backup ID is required")
+		return
+	}
+
+	var req RestoreBackupRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, "password is required for restore operation")
+		return
+	}
+
+	// 从上下文获取当前管理员用户 ID
+	sub, ok := middleware.GetAuthSubjectFromContext(c)
+	if !ok {
+		response.Unauthorized(c, "unauthorized")
+		return
+	}
+
+	// 获取管理员用户并验证密码
+	user, err := h.userService.GetByID(c.Request.Context(), sub.UserID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	if !user.CheckPassword(req.Password) {
+		response.BadRequest(c, "incorrect admin password")
+		return
+	}
+
+	record, err := h.backupService.StartRestore(c.Request.Context(), backupID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+	response.Accepted(c, record)
+}

backend/internal/handler/admin/dashboard_handler.go

@@ -9,6 +9,7 @@ import (
 
 	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
+	"github.com/Wei-Shaw/sub2api/internal/pkg/usagestats"
 	"github.com/Wei-Shaw/sub2api/internal/service"
 
 	"github.com/gin-gonic/gin"
@@ -466,9 +467,62 @@ type BatchUsersUsageRequest struct {
 	UserIDs []int64 `json:"user_ids" binding:"required"`
 }
 
+var dashboardUsersRankingCache = newSnapshotCache(5 * time.Minute)
 var dashboardBatchUsersUsageCache = newSnapshotCache(30 * time.Second)
 var dashboardBatchAPIKeysUsageCache = newSnapshotCache(30 * time.Second)
 
+func parseRankingLimit(raw string) int {
+	limit, err := strconv.Atoi(strings.TrimSpace(raw))
+	if err != nil || limit <= 0 {
+		return 12
+	}
+	if limit > 50 {
+		return 50
+	}
+	return limit
+}
+
+// GetUserSpendingRanking handles getting user spending ranking data.
+// GET /api/v1/admin/dashboard/users-ranking
+func (h *DashboardHandler) GetUserSpendingRanking(c *gin.Context) {
+	startTime, endTime := parseTimeRange(c)
+	limit := parseRankingLimit(c.DefaultQuery("limit", "12"))
+
+	keyRaw, _ := json.Marshal(struct {
+		Start string `json:"start"`
+		End   string `json:"end"`
+		Limit int    `json:"limit"`
+	}{
+		Start: startTime.UTC().Format(time.RFC3339),
+		End:   endTime.UTC().Format(time.RFC3339),
+		Limit: limit,
+	})
+	cacheKey := string(keyRaw)
+	if cached, ok := dashboardUsersRankingCache.Get(cacheKey); ok {
+		c.Header("X-Snapshot-Cache", "hit")
+		response.Success(c, cached.Payload)
+		return
+	}
+
+	ranking, err := h.dashboardService.GetUserSpendingRanking(c.Request.Context(), startTime, endTime, limit)
+	if err != nil {
+		response.Error(c, 500, "Failed to get user spending ranking")
+		return
+	}
+
+	payload := gin.H{
+		"ranking":           ranking.Ranking,
+		"total_actual_cost": ranking.TotalActualCost,
+		"total_requests":    ranking.TotalRequests,
+		"total_tokens":      ranking.TotalTokens,
+		"start_date":        startTime.Format("2006-01-02"),
+		"end_date":          endTime.Add(-24 * time.Hour).Format("2006-01-02"),
+	}
+	dashboardUsersRankingCache.Set(cacheKey, payload)
+	c.Header("X-Snapshot-Cache", "miss")
+	response.Success(c, payload)
+}
+
 // GetBatchUsersUsage handles getting usage stats for multiple users
 // POST /api/v1/admin/dashboard/users-usage
 func (h *DashboardHandler) GetBatchUsersUsage(c *gin.Context) {
@@ -551,3 +605,41 @@ func (h *DashboardHandler) GetBatchAPIKeysUsage(c *gin.Context) {
 	c.Header("X-Snapshot-Cache", "miss")
 	response.Success(c, payload)
 }
+
+// GetUserBreakdown handles getting per-user usage breakdown within a dimension.
+// GET /api/v1/admin/dashboard/user-breakdown
+// Query params: start_date, end_date, group_id, model, endpoint, endpoint_type, limit
+func (h *DashboardHandler) GetUserBreakdown(c *gin.Context) {
+	startTime, endTime := parseTimeRange(c)
+
+	dim := usagestats.UserBreakdownDimension{}
+	if v := c.Query("group_id"); v != "" {
+		if id, err := strconv.ParseInt(v, 10, 64); err == nil {
+			dim.GroupID = id
+		}
+	}
+	dim.Model = c.Query("model")
+	dim.Endpoint = c.Query("endpoint")
+	dim.EndpointType = c.DefaultQuery("endpoint_type", "inbound")
+
+	limit := 50
+	if v := c.Query("limit"); v != "" {
+		if n, err := strconv.Atoi(v); err == nil && n > 0 && n <= 200 {
+			limit = n
+		}
+	}
+
+	stats, err := h.dashboardService.GetUserBreakdownStats(
+		c.Request.Context(), startTime, endTime, dim, limit,
+	)
+	if err != nil {
+		response.Error(c, 500, "Failed to get user breakdown stats")
+		return
+	}
+
+	response.Success(c, gin.H{
+		"users":      stats,
+		"start_date": startTime.Format("2006-01-02"),
+		"end_date":   endTime.Add(-24 * time.Hour).Format("2006-01-02"),
+	})
+}

backend/internal/handler/admin/dashboard_handler_request_type_test.go

@@ -19,6 +19,9 @@ type dashboardUsageRepoCapture struct {
 	trendStream      *bool
 	modelRequestType *int16
 	modelStream      *bool
+	rankingLimit     int
+	ranking          []usagestats.UserSpendingRankingItem
+	rankingTotal     float64
 }
 
 func (s *dashboardUsageRepoCapture) GetUsageTrendWithFilters(
@@ -49,6 +52,20 @@ func (s *dashboardUsageRepoCapture) GetModelStatsWithFilters(
 	return []usagestats.ModelStat{}, nil
 }
 
+func (s *dashboardUsageRepoCapture) GetUserSpendingRanking(
+	ctx context.Context,
+	startTime, endTime time.Time,
+	limit int,
+) (*usagestats.UserSpendingRankingResponse, error) {
+	s.rankingLimit = limit
+	return &usagestats.UserSpendingRankingResponse{
+		Ranking:         s.ranking,
+		TotalActualCost: s.rankingTotal,
+		TotalRequests:   44,
+		TotalTokens:     1234,
+	}, nil
+}
+
 func newDashboardRequestTypeTestRouter(repo *dashboardUsageRepoCapture) *gin.Engine {
 	gin.SetMode(gin.TestMode)
 	dashboardSvc := service.NewDashboardService(repo, nil, nil, nil)
@@ -56,6 +73,7 @@ func newDashboardRequestTypeTestRouter(repo *dashboardUsageRepoCapture) *gin.Eng
 	router := gin.New()
 	router.GET("/admin/dashboard/trend", handler.GetUsageTrend)
 	router.GET("/admin/dashboard/models", handler.GetModelStats)
+	router.GET("/admin/dashboard/users-ranking", handler.GetUserSpendingRanking)
 	return router
 }
 
@@ -130,3 +148,32 @@ func TestDashboardModelStatsInvalidStream(t *testing.T) {
 
 	require.Equal(t, http.StatusBadRequest, rec.Code)
 }
+
+func TestDashboardUsersRankingLimitAndCache(t *testing.T) {
+	dashboardUsersRankingCache = newSnapshotCache(5 * time.Minute)
+	repo := &dashboardUsageRepoCapture{
+		ranking: []usagestats.UserSpendingRankingItem{
+			{UserID: 7, Email: "rank@example.com", ActualCost: 10.5, Requests: 3, Tokens: 300},
+		},
+		rankingTotal: 88.8,
+	}
+	router := newDashboardRequestTypeTestRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet, "/admin/dashboard/users-ranking?limit=100&start_date=2025-01-01&end_date=2025-01-02", nil)
+	rec := httptest.NewRecorder()
+	router.ServeHTTP(rec, req)
+
+	require.Equal(t, http.StatusOK, rec.Code)
+	require.Equal(t, 50, repo.rankingLimit)
+	require.Contains(t, rec.Body.String(), "\"total_actual_cost\":88.8")
+	require.Contains(t, rec.Body.String(), "\"total_requests\":44")
+	require.Contains(t, rec.Body.String(), "\"total_tokens\":1234")
+	require.Equal(t, "miss", rec.Header().Get("X-Snapshot-Cache"))
+
+	req2 := httptest.NewRequest(http.MethodGet, "/admin/dashboard/users-ranking?limit=100&start_date=2025-01-01&end_date=2025-01-02", nil)
+	rec2 := httptest.NewRecorder()
+	router.ServeHTTP(rec2, req2)
+
+	require.Equal(t, http.StatusOK, rec2.Code)
+	require.Equal(t, "hit", rec2.Header().Get("X-Snapshot-Cache"))
+}

backend/internal/handler/admin/dashboard_handler_user_breakdown_test.go

@@ -0,0 +1,203 @@
+package admin
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+	"time"
+
+	"github.com/Wei-Shaw/sub2api/internal/pkg/usagestats"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/require"
+)
+
+// --- mock repo ---
+
+type userBreakdownRepoCapture struct {
+	service.UsageLogRepository
+	capturedDim   usagestats.UserBreakdownDimension
+	capturedLimit int
+	result        []usagestats.UserBreakdownItem
+}
+
+func (r *userBreakdownRepoCapture) GetUserBreakdownStats(
+	_ context.Context, _, _ time.Time,
+	dim usagestats.UserBreakdownDimension, limit int,
+) ([]usagestats.UserBreakdownItem, error) {
+	r.capturedDim = dim
+	r.capturedLimit = limit
+	if r.result != nil {
+		return r.result, nil
+	}
+	return []usagestats.UserBreakdownItem{}, nil
+}
+
+func newUserBreakdownRouter(repo *userBreakdownRepoCapture) *gin.Engine {
+	gin.SetMode(gin.TestMode)
+	svc := service.NewDashboardService(repo, nil, nil, nil)
+	h := NewDashboardHandler(svc, nil)
+	router := gin.New()
+	router.GET("/admin/dashboard/user-breakdown", h.GetUserBreakdown)
+	return router
+}
+
+// --- tests ---
+
+func TestGetUserBreakdown_GroupIDFilter(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&group_id=42", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, int64(42), repo.capturedDim.GroupID)
+	require.Empty(t, repo.capturedDim.Model)
+	require.Empty(t, repo.capturedDim.Endpoint)
+	require.Equal(t, 50, repo.capturedLimit) // default limit
+}
+
+func TestGetUserBreakdown_ModelFilter(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&model=claude-opus-4-6", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, "claude-opus-4-6", repo.capturedDim.Model)
+	require.Equal(t, int64(0), repo.capturedDim.GroupID)
+}
+
+func TestGetUserBreakdown_EndpointFilter(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&endpoint=/v1/messages&endpoint_type=upstream", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, "/v1/messages", repo.capturedDim.Endpoint)
+	require.Equal(t, "upstream", repo.capturedDim.EndpointType)
+}
+
+func TestGetUserBreakdown_DefaultEndpointType(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&endpoint=/chat", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, "inbound", repo.capturedDim.EndpointType)
+}
+
+func TestGetUserBreakdown_CustomLimit(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&model=test&limit=100", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, 100, repo.capturedLimit)
+}
+
+func TestGetUserBreakdown_LimitClamped(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	// limit > 200 should fall back to default 50
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&model=test&limit=999", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, 50, repo.capturedLimit)
+}
+
+func TestGetUserBreakdown_ResponseFormat(t *testing.T) {
+	repo := &userBreakdownRepoCapture{
+		result: []usagestats.UserBreakdownItem{
+			{UserID: 1, Email: "alice@test.com", Requests: 100, TotalTokens: 50000, Cost: 1.5, ActualCost: 1.2},
+			{UserID: 2, Email: "bob@test.com", Requests: 50, TotalTokens: 25000, Cost: 0.8, ActualCost: 0.6},
+		},
+	}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&group_id=1", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+
+	var resp struct {
+		Code int `json:"code"`
+		Data struct {
+			Users     []usagestats.UserBreakdownItem `json:"users"`
+			StartDate string                         `json:"start_date"`
+			EndDate   string                         `json:"end_date"`
+		} `json:"data"`
+	}
+	err := json.Unmarshal(w.Body.Bytes(), &resp)
+	require.NoError(t, err)
+	require.Equal(t, 0, resp.Code)
+	require.Len(t, resp.Data.Users, 2)
+	require.Equal(t, int64(1), resp.Data.Users[0].UserID)
+	require.Equal(t, "alice@test.com", resp.Data.Users[0].Email)
+	require.Equal(t, int64(100), resp.Data.Users[0].Requests)
+	require.InDelta(t, 1.2, resp.Data.Users[0].ActualCost, 0.001)
+	require.Equal(t, "2026-03-01", resp.Data.StartDate)
+	require.Equal(t, "2026-03-16", resp.Data.EndDate)
+}
+
+func TestGetUserBreakdown_EmptyResult(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16&group_id=999", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+
+	var resp struct {
+		Data struct {
+			Users []usagestats.UserBreakdownItem `json:"users"`
+		} `json:"data"`
+	}
+	err := json.Unmarshal(w.Body.Bytes(), &resp)
+	require.NoError(t, err)
+	require.Empty(t, resp.Data.Users)
+}
+
+func TestGetUserBreakdown_NoFilters(t *testing.T) {
+	repo := &userBreakdownRepoCapture{}
+	router := newUserBreakdownRouter(repo)
+
+	req := httptest.NewRequest(http.MethodGet,
+		"/admin/dashboard/user-breakdown?start_date=2026-03-01&end_date=2026-03-16", nil)
+	w := httptest.NewRecorder()
+	router.ServeHTTP(w, req)
+
+	require.Equal(t, http.StatusOK, w.Code)
+	require.Equal(t, int64(0), repo.capturedDim.GroupID)
+	require.Empty(t, repo.capturedDim.Model)
+	require.Empty(t, repo.capturedDim.Endpoint)
+}

backend/internal/handler/admin/group_handler.go

@@ -1,11 +1,15 @@
 package admin
 
 import (
+	"bytes"
+	"encoding/json"
+	"fmt"
 	"strconv"
 	"strings"
 
 	"github.com/Wei-Shaw/sub2api/internal/handler/dto"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
+	"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
 	"github.com/Wei-Shaw/sub2api/internal/service"
 
 	"github.com/gin-gonic/gin"
@@ -13,27 +17,80 @@ import (
 
 // GroupHandler handles admin group management
 type GroupHandler struct {
-	adminService service.AdminService
+	adminService         service.AdminService
+	dashboardService     *service.DashboardService
+	groupCapacityService *service.GroupCapacityService
+}
+
+type optionalLimitField struct {
+	set   bool
+	value *float64
+}
+
+func (f *optionalLimitField) UnmarshalJSON(data []byte) error {
+	f.set = true
+
+	trimmed := bytes.TrimSpace(data)
+	if bytes.Equal(trimmed, []byte("null")) {
+		f.value = nil
+		return nil
+	}
+
+	var number float64
+	if err := json.Unmarshal(trimmed, &number); err == nil {
+		f.value = &number
+		return nil
+	}
+
+	var text string
+	if err := json.Unmarshal(trimmed, &text); err == nil {
+		text = strings.TrimSpace(text)
+		if text == "" {
+			f.value = nil
+			return nil
+		}
+		number, err = strconv.ParseFloat(text, 64)
+		if err != nil {
+			return fmt.Errorf("invalid numeric limit value %q: %w", text, err)
+		}
+		f.value = &number
+		return nil
+	}
+
+	return fmt.Errorf("invalid limit value: %s", string(trimmed))
+}
+
+func (f optionalLimitField) ToServiceInput() *float64 {
+	if !f.set {
+		return nil
+	}
+	if f.value != nil {
+		return f.value
+	}
+	zero := 0.0
+	return &zero
 }
 
 // NewGroupHandler creates a new admin group handler
-func NewGroupHandler(adminService service.AdminService) *GroupHandler {
+func NewGroupHandler(adminService service.AdminService, dashboardService *service.DashboardService, groupCapacityService *service.GroupCapacityService) *GroupHandler {
 	return &GroupHandler{
-		adminService: adminService,
+		adminService:         adminService,
+		dashboardService:     dashboardService,
+		groupCapacityService: groupCapacityService,
 	}
 }
 
 // CreateGroupRequest represents create group request
 type CreateGroupRequest struct {
-	Name             string   `json:"name" binding:"required"`
-	Description      string   `json:"description"`
-	Platform         string   `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity sora"`
-	RateMultiplier   float64  `json:"rate_multiplier"`
-	IsExclusive      bool     `json:"is_exclusive"`
-	SubscriptionType string   `json:"subscription_type" binding:"omitempty,oneof=standard subscription"`
-	DailyLimitUSD    *float64 `json:"daily_limit_usd"`
-	WeeklyLimitUSD   *float64 `json:"weekly_limit_usd"`
-	MonthlyLimitUSD  *float64 `json:"monthly_limit_usd"`
+	Name             string             `json:"name" binding:"required"`
+	Description      string             `json:"description"`
+	Platform         string             `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity sora"`
+	RateMultiplier   float64            `json:"rate_multiplier"`
+	IsExclusive      bool               `json:"is_exclusive"`
+	SubscriptionType string             `json:"subscription_type" binding:"omitempty,oneof=standard subscription"`
+	DailyLimitUSD    optionalLimitField `json:"daily_limit_usd"`
+	WeeklyLimitUSD   optionalLimitField `json:"weekly_limit_usd"`
+	MonthlyLimitUSD  optionalLimitField `json:"monthly_limit_usd"`
 	// 图片生成计费配置（antigravity 和 gemini 平台使用，负数表示清除配置）
 	ImagePrice1K                    *float64 `json:"image_price_1k"`
 	ImagePrice2K                    *float64 `json:"image_price_2k"`
@@ -62,16 +119,16 @@ type CreateGroupRequest struct {
 
 // UpdateGroupRequest represents update group request
 type UpdateGroupRequest struct {
-	Name             string   `json:"name"`
-	Description      string   `json:"description"`
-	Platform         string   `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity sora"`
-	RateMultiplier   *float64 `json:"rate_multiplier"`
-	IsExclusive      *bool    `json:"is_exclusive"`
-	Status           string   `json:"status" binding:"omitempty,oneof=active inactive"`
-	SubscriptionType string   `json:"subscription_type" binding:"omitempty,oneof=standard subscription"`
-	DailyLimitUSD    *float64 `json:"daily_limit_usd"`
-	WeeklyLimitUSD   *float64 `json:"weekly_limit_usd"`
-	MonthlyLimitUSD  *float64 `json:"monthly_limit_usd"`
+	Name             string             `json:"name"`
+	Description      string             `json:"description"`
+	Platform         string             `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity sora"`
+	RateMultiplier   *float64           `json:"rate_multiplier"`
+	IsExclusive      *bool              `json:"is_exclusive"`
+	Status           string             `json:"status" binding:"omitempty,oneof=active inactive"`
+	SubscriptionType string             `json:"subscription_type" binding:"omitempty,oneof=standard subscription"`
+	DailyLimitUSD    optionalLimitField `json:"daily_limit_usd"`
+	WeeklyLimitUSD   optionalLimitField `json:"weekly_limit_usd"`
+	MonthlyLimitUSD  optionalLimitField `json:"monthly_limit_usd"`
 	// 图片生成计费配置（antigravity 和 gemini 平台使用，负数表示清除配置）
 	ImagePrice1K                    *float64 `json:"image_price_1k"`
 	ImagePrice2K                    *float64 `json:"image_price_2k"`
@@ -191,9 +248,9 @@ func (h *GroupHandler) Create(c *gin.Context) {
 		RateMultiplier:                  req.RateMultiplier,
 		IsExclusive:                     req.IsExclusive,
 		SubscriptionType:                req.SubscriptionType,
-		DailyLimitUSD:                   req.DailyLimitUSD,
-		WeeklyLimitUSD:                  req.WeeklyLimitUSD,
-		MonthlyLimitUSD:                 req.MonthlyLimitUSD,
+		DailyLimitUSD:                   req.DailyLimitUSD.ToServiceInput(),
+		WeeklyLimitUSD:                  req.WeeklyLimitUSD.ToServiceInput(),
+		MonthlyLimitUSD:                 req.MonthlyLimitUSD.ToServiceInput(),
 		ImagePrice1K:                    req.ImagePrice1K,
 		ImagePrice2K:                    req.ImagePrice2K,
 		ImagePrice4K:                    req.ImagePrice4K,
@@ -244,9 +301,9 @@ func (h *GroupHandler) Update(c *gin.Context) {
 		IsExclusive:                     req.IsExclusive,
 		Status:                          req.Status,
 		SubscriptionType:                req.SubscriptionType,
-		DailyLimitUSD:                   req.DailyLimitUSD,
-		WeeklyLimitUSD:                  req.WeeklyLimitUSD,
-		MonthlyLimitUSD:                 req.MonthlyLimitUSD,
+		DailyLimitUSD:                   req.DailyLimitUSD.ToServiceInput(),
+		WeeklyLimitUSD:                  req.WeeklyLimitUSD.ToServiceInput(),
+		MonthlyLimitUSD:                 req.MonthlyLimitUSD.ToServiceInput(),
 		ImagePrice1K:                    req.ImagePrice1K,
 		ImagePrice2K:                    req.ImagePrice2K,
 		ImagePrice4K:                    req.ImagePrice4K,
@@ -311,6 +368,33 @@ func (h *GroupHandler) GetStats(c *gin.Context) {
 	_ = groupID // TODO: implement actual stats
 }
 
+// GetUsageSummary returns today's and cumulative cost for all groups.
+// GET /api/v1/admin/groups/usage-summary?timezone=Asia/Shanghai
+func (h *GroupHandler) GetUsageSummary(c *gin.Context) {
+	userTZ := c.Query("timezone")
+	now := timezone.NowInUserLocation(userTZ)
+	todayStart := timezone.StartOfDayInUserLocation(now, userTZ)
+
+	results, err := h.dashboardService.GetGroupUsageSummary(c.Request.Context(), todayStart)
+	if err != nil {
+		response.Error(c, 500, "Failed to get group usage summary")
+		return
+	}
+
+	response.Success(c, results)
+}
+
+// GetCapacitySummary returns aggregated capacity (concurrency/sessions/RPM) for all active groups.
+// GET /api/v1/admin/groups/capacity-summary
+func (h *GroupHandler) GetCapacitySummary(c *gin.Context) {
+	results, err := h.groupCapacityService.GetAllGroupCapacity(c.Request.Context())
+	if err != nil {
+		response.Error(c, 500, "Failed to get group capacity summary")
+		return
+	}
+	response.Success(c, results)
+}
+
 // GetGroupAPIKeys handles getting API keys in a group
 // GET /api/v1/admin/groups/:id/api-keys
 func (h *GroupHandler) GetGroupAPIKeys(c *gin.Context) {
@@ -335,6 +419,72 @@ func (h *GroupHandler) GetGroupAPIKeys(c *gin.Context) {
 	response.Paginated(c, outKeys, total, page, pageSize)
 }
 
+// GetGroupRateMultipliers handles getting rate multipliers for users in a group
+// GET /api/v1/admin/groups/:id/rate-multipliers
+func (h *GroupHandler) GetGroupRateMultipliers(c *gin.Context) {
+	groupID, err := strconv.ParseInt(c.Param("id"), 10, 64)
+	if err != nil {
+		response.BadRequest(c, "Invalid group ID")
+		return
+	}
+
+	entries, err := h.adminService.GetGroupRateMultipliers(c.Request.Context(), groupID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+
+	if entries == nil {
+		entries = []service.UserGroupRateEntry{}
+	}
+	response.Success(c, entries)
+}
+
+// ClearGroupRateMultipliers handles clearing all rate multipliers for a group
+// DELETE /api/v1/admin/groups/:id/rate-multipliers
+func (h *GroupHandler) ClearGroupRateMultipliers(c *gin.Context) {
+	groupID, err := strconv.ParseInt(c.Param("id"), 10, 64)
+	if err != nil {
+		response.BadRequest(c, "Invalid group ID")
+		return
+	}
+
+	if err := h.adminService.ClearGroupRateMultipliers(c.Request.Context(), groupID); err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+
+	response.Success(c, gin.H{"message": "Rate multipliers cleared successfully"})
+}
+
+// BatchSetGroupRateMultipliersRequest represents batch set rate multipliers request
+type BatchSetGroupRateMultipliersRequest struct {
+	Entries []service.GroupRateMultiplierInput `json:"entries" binding:"required"`
+}
+
+// BatchSetGroupRateMultipliers handles batch setting rate multipliers for a group
+// PUT /api/v1/admin/groups/:id/rate-multipliers
+func (h *GroupHandler) BatchSetGroupRateMultipliers(c *gin.Context) {
+	groupID, err := strconv.ParseInt(c.Param("id"), 10, 64)
+	if err != nil {
+		response.BadRequest(c, "Invalid group ID")
+		return
+	}
+
+	var req BatchSetGroupRateMultipliersRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, "Invalid request: "+err.Error())
+		return
+	}
+
+	if err := h.adminService.BatchSetGroupRateMultipliers(c.Request.Context(), groupID, req.Entries); err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+
+	response.Success(c, gin.H{"message": "Rate multipliers updated successfully"})
+}
+
 // UpdateSortOrderRequest represents the request to update group sort orders
 type UpdateSortOrderRequest struct {
 	Updates []struct {

backend/internal/handler/admin/openai_oauth_handler.go

@@ -289,6 +289,7 @@ func (h *OpenAIOAuthHandler) CreateAccountFromOAuth(c *gin.Context) {
 		Platform:    platform,
 		Type:        "oauth",
 		Credentials: credentials,
+		Extra:       nil,
 		ProxyID:     req.ProxyID,
 		Concurrency: req.Concurrency,
 		Priority:    req.Priority,

backend/internal/handler/admin/redeem_handler.go

@@ -41,12 +41,15 @@ type GenerateRedeemCodesRequest struct {
 }
 
 // CreateAndRedeemCodeRequest represents creating a fixed code and redeeming it for a target user.
+// Type 为 omitempty 而非 required 是为了向后兼容旧版调用方（不传 type 时默认 balance）。
 type CreateAndRedeemCodeRequest struct {
-	Code   string  `json:"code" binding:"required,min=3,max=128"`
-	Type   string  `json:"type" binding:"required,oneof=balance concurrency subscription invitation"`
-	Value  float64 `json:"value" binding:"required,gt=0"`
-	UserID int64   `json:"user_id" binding:"required,gt=0"`
-	Notes  string  `json:"notes"`
+	Code         string  `json:"code" binding:"required,min=3,max=128"`
+	Type         string  `json:"type" binding:"omitempty,oneof=balance concurrency subscription invitation"` // 不传时默认 balance（向后兼容）
+	Value        float64 `json:"value" binding:"required,gt=0"`
+	UserID       int64   `json:"user_id" binding:"required,gt=0"`
+	GroupID      *int64  `json:"group_id"`                                    // subscription 类型必填
+	ValidityDays int     `json:"validity_days" binding:"omitempty,max=36500"` // subscription 类型必填，>0
+	Notes        string  `json:"notes"`
 }
 
 // List handles listing all redeem codes with pagination
@@ -136,6 +139,22 @@ func (h *RedeemHandler) CreateAndRedeem(c *gin.Context) {
 		return
 	}
 	req.Code = strings.TrimSpace(req.Code)
+	// 向后兼容：旧版调用方（如 Sub2ApiPay）不传 type 字段，默认当作 balance 充值处理。
+	// 请勿删除此默认值逻辑，否则会导致旧版调用方 400 报错。
+	if req.Type == "" {
+		req.Type = "balance"
+	}
+
+	if req.Type == "subscription" {
+		if req.GroupID == nil {
+			response.BadRequest(c, "group_id is required for subscription type")
+			return
+		}
+		if req.ValidityDays <= 0 {
+			response.BadRequest(c, "validity_days must be greater than 0 for subscription type")
+			return
+		}
+	}
 
 	executeAdminIdempotentJSON(c, "admin.redeem_codes.create_and_redeem", req, service.DefaultWriteIdempotencyTTL(), func(ctx context.Context) (any, error) {
 		existing, err := h.redeemService.GetByCode(ctx, req.Code)
@@ -147,11 +166,13 @@ func (h *RedeemHandler) CreateAndRedeem(c *gin.Context) {
 		}
 
 		createErr := h.redeemService.CreateCode(ctx, &service.RedeemCode{
-			Code:   req.Code,
-			Type:   req.Type,
-			Value:  req.Value,
-			Status: service.StatusUnused,
-			Notes:  req.Notes,
+			Code:         req.Code,
+			Type:         req.Type,
+			Value:        req.Value,
+			Status:       service.StatusUnused,
+			Notes:        req.Notes,
+			GroupID:      req.GroupID,
+			ValidityDays: req.ValidityDays,
 		})
 		if createErr != nil {
 			// Unique code race: if code now exists, use idempotent semantics by used_by.

backend/internal/handler/admin/redeem_handler_test.go

@@ -0,0 +1,135 @@
+package admin
+
+import (
+	"bytes"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// newCreateAndRedeemHandler creates a RedeemHandler with a non-nil (but minimal)
+// RedeemService so that CreateAndRedeem's nil guard passes and we can test the
+// parameter-validation layer that runs before any service call.
+func newCreateAndRedeemHandler() *RedeemHandler {
+	return &RedeemHandler{
+		adminService:  newStubAdminService(),
+		redeemService: &service.RedeemService{}, // non-nil to pass nil guard
+	}
+}
+
+// postCreateAndRedeemValidation calls CreateAndRedeem and returns the response
+// status code. For cases that pass validation and proceed into the service layer,
+// a panic may occur (because RedeemService internals are nil); this is expected
+// and treated as "validation passed" (returns 0 to indicate panic).
+func postCreateAndRedeemValidation(t *testing.T, handler *RedeemHandler, body any) (code int) {
+	t.Helper()
+	gin.SetMode(gin.TestMode)
+	w := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(w)
+
+	jsonBytes, err := json.Marshal(body)
+	require.NoError(t, err)
+	c.Request, _ = http.NewRequest(http.MethodPost, "/api/v1/admin/redeem-codes/create-and-redeem", bytes.NewReader(jsonBytes))
+	c.Request.Header.Set("Content-Type", "application/json")
+
+	defer func() {
+		if r := recover(); r != nil {
+			// Panic means we passed validation and entered service layer (expected for minimal stub).
+			code = 0
+		}
+	}()
+	handler.CreateAndRedeem(c)
+	return w.Code
+}
+
+func TestCreateAndRedeem_TypeDefaultsToBalance(t *testing.T) {
+	// 不传 type 字段时应默认 balance，不触发 subscription 校验。
+	// 验证通过后进入 service 层会 panic（返回 0），说明默认值生效。
+	h := newCreateAndRedeemHandler()
+	code := postCreateAndRedeemValidation(t, h, map[string]any{
+		"code":    "test-balance-default",
+		"value":   10.0,
+		"user_id": 1,
+	})
+
+	assert.NotEqual(t, http.StatusBadRequest, code,
+		"omitting type should default to balance and pass validation")
+}
+
+func TestCreateAndRedeem_SubscriptionRequiresGroupID(t *testing.T) {
+	h := newCreateAndRedeemHandler()
+	code := postCreateAndRedeemValidation(t, h, map[string]any{
+		"code":          "test-sub-no-group",
+		"type":          "subscription",
+		"value":         29.9,
+		"user_id":       1,
+		"validity_days": 30,
+		// group_id 缺失
+	})
+
+	assert.Equal(t, http.StatusBadRequest, code)
+}
+
+func TestCreateAndRedeem_SubscriptionRequiresPositiveValidityDays(t *testing.T) {
+	groupID := int64(5)
+	h := newCreateAndRedeemHandler()
+
+	cases := []struct {
+		name         string
+		validityDays int
+	}{
+		{"zero", 0},
+		{"negative", -1},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			code := postCreateAndRedeemValidation(t, h, map[string]any{
+				"code":          "test-sub-bad-days-" + tc.name,
+				"type":          "subscription",
+				"value":         29.9,
+				"user_id":       1,
+				"group_id":      groupID,
+				"validity_days": tc.validityDays,
+			})
+
+			assert.Equal(t, http.StatusBadRequest, code)
+		})
+	}
+}
+
+func TestCreateAndRedeem_SubscriptionValidParamsPassValidation(t *testing.T) {
+	groupID := int64(5)
+	h := newCreateAndRedeemHandler()
+	code := postCreateAndRedeemValidation(t, h, map[string]any{
+		"code":          "test-sub-valid",
+		"type":          "subscription",
+		"value":         29.9,
+		"user_id":       1,
+		"group_id":      groupID,
+		"validity_days": 31,
+	})
+
+	assert.NotEqual(t, http.StatusBadRequest, code,
+		"valid subscription params should pass validation")
+}
+
+func TestCreateAndRedeem_BalanceIgnoresSubscriptionFields(t *testing.T) {
+	h := newCreateAndRedeemHandler()
+	// balance 类型不传 group_id 和 validity_days，不应报 400
+	code := postCreateAndRedeemValidation(t, h, map[string]any{
+		"code":    "test-balance-no-extras",
+		"type":    "balance",
+		"value":   50.0,
+		"user_id": 1,
+	})
+
+	assert.NotEqual(t, http.StatusBadRequest, code,
+		"balance type should not require group_id or validity_days")
+}

backend/internal/handler/admin/setting_handler.go

@@ -80,6 +80,7 @@ func (h *SettingHandler) GetSettings(c *gin.Context) {
 		RegistrationEmailSuffixWhitelist:     settings.RegistrationEmailSuffixWhitelist,
 		PromoCodeEnabled:                     settings.PromoCodeEnabled,
 		PasswordResetEnabled:                 settings.PasswordResetEnabled,
+		FrontendURL:                          settings.FrontendURL,
 		InvitationCodeEnabled:                settings.InvitationCodeEnabled,
 		TotpEnabled:                          settings.TotpEnabled,
 		TotpEncryptionKeyConfigured:          h.settingService.IsTotpEncryptionKeyConfigured(),
@@ -125,6 +126,7 @@ func (h *SettingHandler) GetSettings(c *gin.Context) {
 		OpsMetricsIntervalSeconds:            settings.OpsMetricsIntervalSeconds,
 		MinClaudeCodeVersion:                 settings.MinClaudeCodeVersion,
 		AllowUngroupedKeyScheduling:          settings.AllowUngroupedKeyScheduling,
+		BackendModeEnabled:                   settings.BackendModeEnabled,
 	})
 }
 
@@ -136,6 +138,7 @@ type UpdateSettingsRequest struct {
 	RegistrationEmailSuffixWhitelist []string `json:"registration_email_suffix_whitelist"`
 	PromoCodeEnabled                 bool     `json:"promo_code_enabled"`
 	PasswordResetEnabled             bool     `json:"password_reset_enabled"`
+	FrontendURL                      string   `json:"frontend_url"`
 	InvitationCodeEnabled            bool     `json:"invitation_code_enabled"`
 	TotpEnabled                      bool     `json:"totp_enabled"` // TOTP 双因素认证
 
@@ -199,6 +202,9 @@ type UpdateSettingsRequest struct {
 
 	// 分组隔离
 	AllowUngroupedKeyScheduling bool `json:"allow_ungrouped_key_scheduling"`
+
+	// Backend Mode
+	BackendModeEnabled bool `json:"backend_mode_enabled"`
 }
 
 // UpdateSettings 更新系统设置
@@ -322,6 +328,15 @@ func (h *SettingHandler) UpdateSettings(c *gin.Context) {
 		}
 	}
 
+	// Frontend URL 验证
+	req.FrontendURL = strings.TrimSpace(req.FrontendURL)
+	if req.FrontendURL != "" {
+		if err := config.ValidateAbsoluteHTTPURL(req.FrontendURL); err != nil {
+			response.BadRequest(c, "Frontend URL must be an absolute http(s) URL")
+			return
+		}
+	}
+
 	// 自定义菜单项验证
 	const (
 		maxCustomMenuItems    = 20
@@ -433,6 +448,7 @@ func (h *SettingHandler) UpdateSettings(c *gin.Context) {
 		RegistrationEmailSuffixWhitelist: req.RegistrationEmailSuffixWhitelist,
 		PromoCodeEnabled:                 req.PromoCodeEnabled,
 		PasswordResetEnabled:             req.PasswordResetEnabled,
+		FrontendURL:                      req.FrontendURL,
 		InvitationCodeEnabled:            req.InvitationCodeEnabled,
 		TotpEnabled:                      req.TotpEnabled,
 		SMTPHost:                         req.SMTPHost,
@@ -473,6 +489,7 @@ func (h *SettingHandler) UpdateSettings(c *gin.Context) {
 		IdentityPatchPrompt:              req.IdentityPatchPrompt,
 		MinClaudeCodeVersion:             req.MinClaudeCodeVersion,
 		AllowUngroupedKeyScheduling:      req.AllowUngroupedKeyScheduling,
+		BackendModeEnabled:               req.BackendModeEnabled,
 		OpsMonitoringEnabled: func() bool {
 			if req.OpsMonitoringEnabled != nil {
 				return *req.OpsMonitoringEnabled
@@ -526,6 +543,7 @@ func (h *SettingHandler) UpdateSettings(c *gin.Context) {
 		RegistrationEmailSuffixWhitelist:     updatedSettings.RegistrationEmailSuffixWhitelist,
 		PromoCodeEnabled:                     updatedSettings.PromoCodeEnabled,
 		PasswordResetEnabled:                 updatedSettings.PasswordResetEnabled,
+		FrontendURL:                          updatedSettings.FrontendURL,
 		InvitationCodeEnabled:                updatedSettings.InvitationCodeEnabled,
 		TotpEnabled:                          updatedSettings.TotpEnabled,
 		TotpEncryptionKeyConfigured:          h.settingService.IsTotpEncryptionKeyConfigured(),
@@ -571,6 +589,7 @@ func (h *SettingHandler) UpdateSettings(c *gin.Context) {
 		OpsMetricsIntervalSeconds:            updatedSettings.OpsMetricsIntervalSeconds,
 		MinClaudeCodeVersion:                 updatedSettings.MinClaudeCodeVersion,
 		AllowUngroupedKeyScheduling:          updatedSettings.AllowUngroupedKeyScheduling,
+		BackendModeEnabled:                   updatedSettings.BackendModeEnabled,
 	})
 }
 
@@ -608,6 +627,9 @@ func diffSettings(before *service.SystemSettings, after *service.SystemSettings,
 	if before.PasswordResetEnabled != after.PasswordResetEnabled {
 		changed = append(changed, "password_reset_enabled")
 	}
+	if before.FrontendURL != after.FrontendURL {
+		changed = append(changed, "frontend_url")
+	}
 	if before.TotpEnabled != after.TotpEnabled {
 		changed = append(changed, "totp_enabled")
 	}
@@ -725,6 +747,9 @@ func diffSettings(before *service.SystemSettings, after *service.SystemSettings,
 	if before.AllowUngroupedKeyScheduling != after.AllowUngroupedKeyScheduling {
 		changed = append(changed, "allow_ungrouped_key_scheduling")
 	}
+	if before.BackendModeEnabled != after.BackendModeEnabled {
+		changed = append(changed, "backend_mode_enabled")
+	}
 	if before.PurchaseSubscriptionEnabled != after.PurchaseSubscriptionEnabled {
 		changed = append(changed, "purchase_subscription_enabled")
 	}

backend/internal/handler/admin/subscription_handler.go

@@ -77,12 +77,13 @@ func (h *SubscriptionHandler) List(c *gin.Context) {
 		}
 	}
 	status := c.Query("status")
+	platform := c.Query("platform")
 
 	// Parse sorting parameters
 	sortBy := c.DefaultQuery("sort_by", "created_at")
 	sortOrder := c.DefaultQuery("sort_order", "desc")
 
-	subscriptions, pagination, err := h.subscriptionService.List(c.Request.Context(), page, pageSize, userID, groupID, status, sortBy, sortOrder)
+	subscriptions, pagination, err := h.subscriptionService.List(c.Request.Context(), page, pageSize, userID, groupID, status, platform, sortBy, sortOrder)
 	if err != nil {
 		response.ErrorFrom(c, err)
 		return
@@ -218,11 +219,12 @@ func (h *SubscriptionHandler) Extend(c *gin.Context) {
 
 // ResetSubscriptionQuotaRequest represents the reset quota request
 type ResetSubscriptionQuotaRequest struct {
-	Daily  bool `json:"daily"`
-	Weekly bool `json:"weekly"`
+	Daily   bool `json:"daily"`
+	Weekly  bool `json:"weekly"`
+	Monthly bool `json:"monthly"`
 }
 
-// ResetQuota resets daily and/or weekly usage for a subscription.
+// ResetQuota resets daily, weekly, and/or monthly usage for a subscription.
 // POST /api/v1/admin/subscriptions/:id/reset-quota
 func (h *SubscriptionHandler) ResetQuota(c *gin.Context) {
 	subscriptionID, err := strconv.ParseInt(c.Param("id"), 10, 64)
@@ -235,11 +237,11 @@ func (h *SubscriptionHandler) ResetQuota(c *gin.Context) {
 		response.BadRequest(c, "Invalid request: "+err.Error())
 		return
 	}
-	if !req.Daily && !req.Weekly {
-		response.BadRequest(c, "At least one of 'daily' or 'weekly' must be true")
+	if !req.Daily && !req.Weekly && !req.Monthly {
+		response.BadRequest(c, "At least one of 'daily', 'weekly', or 'monthly' must be true")
 		return
 	}
-	sub, err := h.subscriptionService.AdminResetQuota(c.Request.Context(), subscriptionID, req.Daily, req.Weekly)
+	sub, err := h.subscriptionService.AdminResetQuota(c.Request.Context(), subscriptionID, req.Daily, req.Weekly, req.Monthly)
 	if err != nil {
 		response.ErrorFrom(c, err)
 		return

backend/internal/handler/admin/usage_handler.go

@@ -159,8 +159,8 @@ func (h *UsageHandler) List(c *gin.Context) {
 			response.BadRequest(c, "Invalid end_date format, use YYYY-MM-DD")
 			return
 		}
-		// Set end time to end of day
-		t = t.Add(24*time.Hour - time.Nanosecond)
+		// Use half-open range [start, end), move to next calendar day start (DST-safe).
+		t = t.AddDate(0, 0, 1)
 		endTime = &t
 	}
 
@@ -285,7 +285,8 @@ func (h *UsageHandler) Stats(c *gin.Context) {
 			response.BadRequest(c, "Invalid end_date format, use YYYY-MM-DD")
 			return
 		}
-		endTime = endTime.Add(24*time.Hour - time.Nanosecond)
+		// 与 SQL 条件 created_at < end 对齐，使用次日 00:00 作为上边界（DST-safe）。
+		endTime = endTime.AddDate(0, 0, 1)
 	} else {
 		period := c.DefaultQuery("period", "today")
 		switch period {

backend/internal/handler/auth_handler.go

@@ -194,6 +194,12 @@ func (h *AuthHandler) Login(c *gin.Context) {
 		return
 	}
 
+	// Backend mode: only admin can login
+	if h.settingSvc.IsBackendModeEnabled(c.Request.Context()) && !user.IsAdmin() {
+		response.Forbidden(c, "Backend mode is active. Only admin login is allowed.")
+		return
+	}
+
 	h.respondWithTokenPair(c, user)
 }
 
@@ -250,16 +256,22 @@ func (h *AuthHandler) Login2FA(c *gin.Context) {
 		return
 	}
 
-	// Delete the login session
-	_ = h.totpService.DeleteLoginSession(c.Request.Context(), req.TempToken)
-
-	// Get the user
+	// Get the user (before session deletion so we can check backend mode)
 	user, err := h.userService.GetByID(c.Request.Context(), session.UserID)
 	if err != nil {
 		response.ErrorFrom(c, err)
 		return
 	}
 
+	// Backend mode: only admin can login (check BEFORE deleting session)
+	if h.settingSvc.IsBackendModeEnabled(c.Request.Context()) && !user.IsAdmin() {
+		response.Forbidden(c, "Backend mode is active. Only admin login is allowed.")
+		return
+	}
+
+	// Delete the login session (only after all checks pass)
+	_ = h.totpService.DeleteLoginSession(c.Request.Context(), req.TempToken)
+
 	h.respondWithTokenPair(c, user)
 }
 
@@ -447,9 +459,9 @@ func (h *AuthHandler) ForgotPassword(c *gin.Context) {
 		return
 	}
 
-	frontendBaseURL := strings.TrimSpace(h.cfg.Server.FrontendURL)
+	frontendBaseURL := strings.TrimSpace(h.settingSvc.GetFrontendURL(c.Request.Context()))
 	if frontendBaseURL == "" {
-		slog.Error("server.frontend_url not configured; cannot build password reset link")
+		slog.Error("frontend_url not configured in settings or config; cannot build password reset link")
 		response.InternalError(c, "Password reset is not configured")
 		return
 	}
@@ -522,16 +534,22 @@ func (h *AuthHandler) RefreshToken(c *gin.Context) {
 		return
 	}
 
-	tokenPair, err := h.authService.RefreshTokenPair(c.Request.Context(), req.RefreshToken)
+	result, err := h.authService.RefreshTokenPair(c.Request.Context(), req.RefreshToken)
 	if err != nil {
 		response.ErrorFrom(c, err)
 		return
 	}
 
+	// Backend mode: block non-admin token refresh
+	if h.settingSvc.IsBackendModeEnabled(c.Request.Context()) && result.UserRole != "admin" {
+		response.Forbidden(c, "Backend mode is active. Only admin login is allowed.")
+		return
+	}
+
 	response.Success(c, RefreshTokenResponse{
-		AccessToken:  tokenPair.AccessToken,
-		RefreshToken: tokenPair.RefreshToken,
-		ExpiresIn:    tokenPair.ExpiresIn,
+		AccessToken:  result.AccessToken,
+		RefreshToken: result.RefreshToken,
+		ExpiresIn:    result.ExpiresIn,
 		TokenType:    "Bearer",
 	})
 }

backend/internal/handler/dto/mappers.go

@@ -135,14 +135,16 @@ func GroupFromServiceAdmin(g *service.Group) *AdminGroup {
 		return nil
 	}
 	out := &AdminGroup{
-		Group:                groupFromServiceBase(g),
-		ModelRouting:         g.ModelRouting,
-		ModelRoutingEnabled:  g.ModelRoutingEnabled,
-		MCPXMLInject:         g.MCPXMLInject,
-		DefaultMappedModel:   g.DefaultMappedModel,
-		SupportedModelScopes: g.SupportedModelScopes,
-		AccountCount:         g.AccountCount,
-		SortOrder:            g.SortOrder,
+		Group:                   groupFromServiceBase(g),
+		ModelRouting:            g.ModelRouting,
+		ModelRoutingEnabled:     g.ModelRoutingEnabled,
+		MCPXMLInject:            g.MCPXMLInject,
+		DefaultMappedModel:      g.DefaultMappedModel,
+		SupportedModelScopes:    g.SupportedModelScopes,
+		AccountCount:            g.AccountCount,
+		ActiveAccountCount:      g.ActiveAccountCount,
+		RateLimitedAccountCount: g.RateLimitedAccountCount,
+		SortOrder:               g.SortOrder,
 	}
 	if len(g.AccountGroups) > 0 {
 		out.AccountGroups = make([]AccountGroup, 0, len(g.AccountGroups))
@@ -264,8 +266,8 @@ func AccountFromServiceShallow(a *service.Account) *Account {
 		}
 	}
 
-	// 提取 API Key 账号配额限制（仅 apikey 类型有效）
-	if a.Type == service.AccountTypeAPIKey {
+	// 提取账号配额限制（apikey / bedrock 类型有效）
+	if a.IsAPIKeyOrBedrock() {
 		if limit := a.GetQuotaLimit(); limit > 0 {
 			out.QuotaLimit = &limit
 			used := a.GetQuotaUsed()
@@ -281,6 +283,31 @@ func AccountFromServiceShallow(a *service.Account) *Account {
 			used := a.GetQuotaWeeklyUsed()
 			out.QuotaWeeklyUsed = &used
 		}
+		// 固定时间重置配置
+		if mode := a.GetQuotaDailyResetMode(); mode == "fixed" {
+			out.QuotaDailyResetMode = &mode
+			hour := a.GetQuotaDailyResetHour()
+			out.QuotaDailyResetHour = &hour
+		}
+		if mode := a.GetQuotaWeeklyResetMode(); mode == "fixed" {
+			out.QuotaWeeklyResetMode = &mode
+			day := a.GetQuotaWeeklyResetDay()
+			out.QuotaWeeklyResetDay = &day
+			hour := a.GetQuotaWeeklyResetHour()
+			out.QuotaWeeklyResetHour = &hour
+		}
+		if a.GetQuotaDailyResetMode() == "fixed" || a.GetQuotaWeeklyResetMode() == "fixed" {
+			tz := a.GetQuotaResetTimezone()
+			out.QuotaResetTimezone = &tz
+		}
+		if a.Extra != nil {
+			if v, ok := a.Extra["quota_daily_reset_at"].(string); ok && v != "" {
+				out.QuotaDailyResetAt = &v
+			}
+			if v, ok := a.Extra["quota_weekly_reset_at"].(string); ok && v != "" {
+				out.QuotaWeeklyResetAt = &v
+			}
+		}
 	}
 
 	return out
@@ -498,6 +525,8 @@ func usageLogFromServiceUser(l *service.UsageLog) UsageLog {
 		Model:                 l.Model,
 		ServiceTier:           l.ServiceTier,
 		ReasoningEffort:       l.ReasoningEffort,
+		InboundEndpoint:       l.InboundEndpoint,
+		UpstreamEndpoint:      l.UpstreamEndpoint,
 		GroupID:               l.GroupID,
 		SubscriptionID:        l.SubscriptionID,
 		InputTokens:           l.InputTokens,

backend/internal/handler/dto/mappers_usage_test.go

@@ -76,10 +76,14 @@ func TestUsageLogFromService_IncludesServiceTierForUserAndAdmin(t *testing.T) {
 	t.Parallel()
 
 	serviceTier := "priority"
+	inboundEndpoint := "/v1/chat/completions"
+	upstreamEndpoint := "/v1/responses"
 	log := &service.UsageLog{
 		RequestID:             "req_3",
 		Model:                 "gpt-5.4",
 		ServiceTier:           &serviceTier,
+		InboundEndpoint:       &inboundEndpoint,
+		UpstreamEndpoint:      &upstreamEndpoint,
 		AccountRateMultiplier: f64Ptr(1.5),
 	}
 
@@ -88,8 +92,16 @@ func TestUsageLogFromService_IncludesServiceTierForUserAndAdmin(t *testing.T) {
 
 	require.NotNil(t, userDTO.ServiceTier)
 	require.Equal(t, serviceTier, *userDTO.ServiceTier)
+	require.NotNil(t, userDTO.InboundEndpoint)
+	require.Equal(t, inboundEndpoint, *userDTO.InboundEndpoint)
+	require.NotNil(t, userDTO.UpstreamEndpoint)
+	require.Equal(t, upstreamEndpoint, *userDTO.UpstreamEndpoint)
 	require.NotNil(t, adminDTO.ServiceTier)
 	require.Equal(t, serviceTier, *adminDTO.ServiceTier)
+	require.NotNil(t, adminDTO.InboundEndpoint)
+	require.Equal(t, inboundEndpoint, *adminDTO.InboundEndpoint)
+	require.NotNil(t, adminDTO.UpstreamEndpoint)
+	require.Equal(t, upstreamEndpoint, *adminDTO.UpstreamEndpoint)
 	require.NotNil(t, adminDTO.AccountRateMultiplier)
 	require.InDelta(t, 1.5, *adminDTO.AccountRateMultiplier, 1e-12)
 }

backend/internal/handler/dto/settings.go

@@ -22,6 +22,7 @@ type SystemSettings struct {
 	RegistrationEmailSuffixWhitelist []string `json:"registration_email_suffix_whitelist"`
 	PromoCodeEnabled                 bool     `json:"promo_code_enabled"`
 	PasswordResetEnabled             bool     `json:"password_reset_enabled"`
+	FrontendURL                      string   `json:"frontend_url"`
 	InvitationCodeEnabled            bool     `json:"invitation_code_enabled"`
 	TotpEnabled                      bool     `json:"totp_enabled"`                   // TOTP 双因素认证
 	TotpEncryptionKeyConfigured      bool     `json:"totp_encryption_key_configured"` // TOTP 加密密钥是否已配置
@@ -81,6 +82,9 @@ type SystemSettings struct {
 
 	// 分组隔离
 	AllowUngroupedKeyScheduling bool `json:"allow_ungrouped_key_scheduling"`
+
+	// Backend Mode
+	BackendModeEnabled bool `json:"backend_mode_enabled"`
 }
 
 type DefaultSubscriptionSetting struct {
@@ -111,6 +115,7 @@ type PublicSettings struct {
 	CustomMenuItems                  []CustomMenuItem `json:"custom_menu_items"`
 	LinuxDoOAuthEnabled              bool             `json:"linuxdo_oauth_enabled"`
 	SoraClientEnabled                bool             `json:"sora_client_enabled"`
+	BackendModeEnabled               bool             `json:"backend_mode_enabled"`
 	Version                          string           `json:"version"`
 }
 

backend/internal/handler/dto/types.go

@@ -122,9 +122,11 @@ type AdminGroup struct {
 	DefaultMappedModel string `json:"default_mapped_model"`
 
 	// 支持的模型系列（仅 antigravity 平台使用）
-	SupportedModelScopes []string       `json:"supported_model_scopes"`
-	AccountGroups        []AccountGroup `json:"account_groups,omitempty"`
-	AccountCount         int64          `json:"account_count,omitempty"`
+	SupportedModelScopes    []string       `json:"supported_model_scopes"`
+	AccountGroups           []AccountGroup `json:"account_groups,omitempty"`
+	AccountCount            int64          `json:"account_count,omitempty"`
+	ActiveAccountCount      int64          `json:"active_account_count,omitempty"`
+	RateLimitedAccountCount int64          `json:"rate_limited_account_count,omitempty"`
 
 	// 分组排序
 	SortOrder int `json:"sort_order"`
@@ -203,6 +205,16 @@ type Account struct {
 	QuotaWeeklyLimit *float64 `json:"quota_weekly_limit,omitempty"`
 	QuotaWeeklyUsed  *float64 `json:"quota_weekly_used,omitempty"`
 
+	// 配额固定时间重置配置
+	QuotaDailyResetMode  *string `json:"quota_daily_reset_mode,omitempty"`
+	QuotaDailyResetHour  *int    `json:"quota_daily_reset_hour,omitempty"`
+	QuotaWeeklyResetMode *string `json:"quota_weekly_reset_mode,omitempty"`
+	QuotaWeeklyResetDay  *int    `json:"quota_weekly_reset_day,omitempty"`
+	QuotaWeeklyResetHour *int    `json:"quota_weekly_reset_hour,omitempty"`
+	QuotaResetTimezone   *string `json:"quota_reset_timezone,omitempty"`
+	QuotaDailyResetAt    *string `json:"quota_daily_reset_at,omitempty"`
+	QuotaWeeklyResetAt   *string `json:"quota_weekly_reset_at,omitempty"`
+
 	Proxy         *Proxy         `json:"proxy,omitempty"`
 	AccountGroups []AccountGroup `json:"account_groups,omitempty"`
 
@@ -324,9 +336,13 @@ type UsageLog struct {
 	Model     string `json:"model"`
 	// ServiceTier records the OpenAI service tier used for billing, e.g. "priority" / "flex".
 	ServiceTier *string `json:"service_tier,omitempty"`
-	// ReasoningEffort is the request's reasoning effort level (OpenAI Responses API).
-	// nil means not provided / not applicable.
+	// ReasoningEffort is the request's reasoning effort level.
+	// OpenAI: "low"/"medium"/"high"/"xhigh"; Claude: "low"/"medium"/"high"/"max".
 	ReasoningEffort *string `json:"reasoning_effort,omitempty"`
+	// InboundEndpoint is the client-facing API endpoint path, e.g. /v1/chat/completions.
+	InboundEndpoint *string `json:"inbound_endpoint,omitempty"`
+	// UpstreamEndpoint is the normalized upstream endpoint path, e.g. /v1/responses.
+	UpstreamEndpoint *string `json:"upstream_endpoint,omitempty"`
 
 	GroupID        *int64 `json:"group_id"`
 	SubscriptionID *int64 `json:"subscription_id"`

backend/internal/handler/endpoint.go

@@ -0,0 +1,174 @@
+package handler
+
+import (
+	"strings"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+)
+
+// ──────────────────────────────────────────────────────────
+// Canonical inbound / upstream endpoint paths.
+// All normalization and derivation reference this single set
+// of constants — add new paths HERE when a new API surface
+// is introduced.
+// ──────────────────────────────────────────────────────────
+
+const (
+	EndpointMessages        = "/v1/messages"
+	EndpointChatCompletions = "/v1/chat/completions"
+	EndpointResponses       = "/v1/responses"
+	EndpointGeminiModels    = "/v1beta/models"
+)
+
+// gin.Context keys used by the middleware and helpers below.
+const (
+	ctxKeyInboundEndpoint = "_gateway_inbound_endpoint"
+)
+
+// ──────────────────────────────────────────────────────────
+// Normalization functions
+// ──────────────────────────────────────────────────────────
+
+// NormalizeInboundEndpoint maps a raw request path (which may carry
+// prefixes like /antigravity, /openai, /sora) to its canonical form.
+//
+//	"/antigravity/v1/messages"   → "/v1/messages"
+//	"/v1/chat/completions"       → "/v1/chat/completions"
+//	"/openai/v1/responses/foo"   → "/v1/responses"
+//	"/v1beta/models/gemini:gen"  → "/v1beta/models"
+func NormalizeInboundEndpoint(path string) string {
+	path = strings.TrimSpace(path)
+	switch {
+	case strings.Contains(path, EndpointChatCompletions):
+		return EndpointChatCompletions
+	case strings.Contains(path, EndpointMessages):
+		return EndpointMessages
+	case strings.Contains(path, EndpointResponses):
+		return EndpointResponses
+	case strings.Contains(path, EndpointGeminiModels):
+		return EndpointGeminiModels
+	default:
+		return path
+	}
+}
+
+// DeriveUpstreamEndpoint determines the upstream endpoint from the
+// account platform and the normalized inbound endpoint.
+//
+// Platform-specific rules:
+//   - OpenAI always forwards to /v1/responses (with optional subpath
+//     such as /v1/responses/compact preserved from the raw URL).
+//   - Anthropic  → /v1/messages
+//   - Gemini     → /v1beta/models
+//   - Sora       → /v1/chat/completions
+//   - Antigravity routes may target either Claude or Gemini, so the
+//     inbound endpoint is used to distinguish.
+func DeriveUpstreamEndpoint(inbound, rawRequestPath, platform string) string {
+	inbound = strings.TrimSpace(inbound)
+
+	switch platform {
+	case service.PlatformOpenAI:
+		// OpenAI forwards everything to the Responses API.
+		// Preserve subresource suffix (e.g. /v1/responses/compact).
+		if suffix := responsesSubpathSuffix(rawRequestPath); suffix != "" {
+			return EndpointResponses + suffix
+		}
+		return EndpointResponses
+
+	case service.PlatformAnthropic:
+		return EndpointMessages
+
+	case service.PlatformGemini:
+		return EndpointGeminiModels
+
+	case service.PlatformSora:
+		return EndpointChatCompletions
+
+	case service.PlatformAntigravity:
+		// Antigravity accounts serve both Claude and Gemini.
+		if inbound == EndpointGeminiModels {
+			return EndpointGeminiModels
+		}
+		return EndpointMessages
+	}
+
+	// Unknown platform — fall back to inbound.
+	return inbound
+}
+
+// responsesSubpathSuffix extracts the part after "/responses" in a raw
+// request path, e.g. "/openai/v1/responses/compact" → "/compact".
+// Returns "" when there is no meaningful suffix.
+func responsesSubpathSuffix(rawPath string) string {
+	trimmed := strings.TrimRight(strings.TrimSpace(rawPath), "/")
+	idx := strings.LastIndex(trimmed, "/responses")
+	if idx < 0 {
+		return ""
+	}
+	suffix := trimmed[idx+len("/responses"):]
+	if suffix == "" || suffix == "/" {
+		return ""
+	}
+	if !strings.HasPrefix(suffix, "/") {
+		return ""
+	}
+	return suffix
+}
+
+// ──────────────────────────────────────────────────────────
+// Middleware
+// ──────────────────────────────────────────────────────────
+
+// InboundEndpointMiddleware normalizes the request path and stores the
+// canonical inbound endpoint in gin.Context so that every handler in
+// the chain can read it via GetInboundEndpoint.
+//
+// Apply this middleware to all gateway route groups.
+func InboundEndpointMiddleware() gin.HandlerFunc {
+	return func(c *gin.Context) {
+		path := c.FullPath()
+		if path == "" && c.Request != nil && c.Request.URL != nil {
+			path = c.Request.URL.Path
+		}
+		c.Set(ctxKeyInboundEndpoint, NormalizeInboundEndpoint(path))
+		c.Next()
+	}
+}
+
+// ──────────────────────────────────────────────────────────
+// Context helpers — used by handlers before building
+// RecordUsageInput / RecordUsageLongContextInput.
+// ──────────────────────────────────────────────────────────
+
+// GetInboundEndpoint returns the canonical inbound endpoint stored by
+// InboundEndpointMiddleware. If the middleware did not run (e.g. in
+// tests), it falls back to normalizing c.FullPath() on the fly.
+func GetInboundEndpoint(c *gin.Context) string {
+	if v, ok := c.Get(ctxKeyInboundEndpoint); ok {
+		if s, ok := v.(string); ok && s != "" {
+			return s
+		}
+	}
+	// Fallback: normalize on the fly.
+	path := ""
+	if c != nil {
+		path = c.FullPath()
+		if path == "" && c.Request != nil && c.Request.URL != nil {
+			path = c.Request.URL.Path
+		}
+	}
+	return NormalizeInboundEndpoint(path)
+}
+
+// GetUpstreamEndpoint derives the upstream endpoint from the context
+// and the account platform. Handlers call this after scheduling an
+// account, passing account.Platform.
+func GetUpstreamEndpoint(c *gin.Context, platform string) string {
+	inbound := GetInboundEndpoint(c)
+	rawPath := ""
+	if c != nil && c.Request != nil && c.Request.URL != nil {
+		rawPath = c.Request.URL.Path
+	}
+	return DeriveUpstreamEndpoint(inbound, rawPath, platform)
+}

backend/internal/handler/endpoint_test.go

@@ -0,0 +1,159 @@
+package handler
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/require"
+)
+
+func init() { gin.SetMode(gin.TestMode) }
+
+// ──────────────────────────────────────────────────────────
+// NormalizeInboundEndpoint
+// ──────────────────────────────────────────────────────────
+
+func TestNormalizeInboundEndpoint(t *testing.T) {
+	tests := []struct {
+		path string
+		want string
+	}{
+		// Direct canonical paths.
+		{"/v1/messages", EndpointMessages},
+		{"/v1/chat/completions", EndpointChatCompletions},
+		{"/v1/responses", EndpointResponses},
+		{"/v1beta/models", EndpointGeminiModels},
+
+		// Prefixed paths (antigravity, openai, sora).
+		{"/antigravity/v1/messages", EndpointMessages},
+		{"/openai/v1/responses", EndpointResponses},
+		{"/openai/v1/responses/compact", EndpointResponses},
+		{"/sora/v1/chat/completions", EndpointChatCompletions},
+		{"/antigravity/v1beta/models/gemini:generateContent", EndpointGeminiModels},
+
+		// Gin route patterns with wildcards.
+		{"/v1beta/models/*modelAction", EndpointGeminiModels},
+		{"/v1/responses/*subpath", EndpointResponses},
+
+		// Unknown path is returned as-is.
+		{"/v1/embeddings", "/v1/embeddings"},
+		{"", ""},
+		{"  /v1/messages  ", EndpointMessages},
+	}
+	for _, tt := range tests {
+		t.Run(tt.path, func(t *testing.T) {
+			require.Equal(t, tt.want, NormalizeInboundEndpoint(tt.path))
+		})
+	}
+}
+
+// ──────────────────────────────────────────────────────────
+// DeriveUpstreamEndpoint
+// ──────────────────────────────────────────────────────────
+
+func TestDeriveUpstreamEndpoint(t *testing.T) {
+	tests := []struct {
+		name     string
+		inbound  string
+		rawPath  string
+		platform string
+		want     string
+	}{
+		// Anthropic.
+		{"anthropic messages", EndpointMessages, "/v1/messages", service.PlatformAnthropic, EndpointMessages},
+
+		// Gemini.
+		{"gemini models", EndpointGeminiModels, "/v1beta/models/gemini:gen", service.PlatformGemini, EndpointGeminiModels},
+
+		// Sora.
+		{"sora completions", EndpointChatCompletions, "/sora/v1/chat/completions", service.PlatformSora, EndpointChatCompletions},
+
+		// OpenAI — always /v1/responses.
+		{"openai responses root", EndpointResponses, "/v1/responses", service.PlatformOpenAI, EndpointResponses},
+		{"openai responses compact", EndpointResponses, "/openai/v1/responses/compact", service.PlatformOpenAI, "/v1/responses/compact"},
+		{"openai responses nested", EndpointResponses, "/openai/v1/responses/compact/detail", service.PlatformOpenAI, "/v1/responses/compact/detail"},
+		{"openai from messages", EndpointMessages, "/v1/messages", service.PlatformOpenAI, EndpointResponses},
+		{"openai from completions", EndpointChatCompletions, "/v1/chat/completions", service.PlatformOpenAI, EndpointResponses},
+
+		// Antigravity — uses inbound to pick Claude vs Gemini upstream.
+		{"antigravity claude", EndpointMessages, "/antigravity/v1/messages", service.PlatformAntigravity, EndpointMessages},
+		{"antigravity gemini", EndpointGeminiModels, "/antigravity/v1beta/models", service.PlatformAntigravity, EndpointGeminiModels},
+
+		// Unknown platform — passthrough.
+		{"unknown platform", "/v1/embeddings", "/v1/embeddings", "unknown", "/v1/embeddings"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			require.Equal(t, tt.want, DeriveUpstreamEndpoint(tt.inbound, tt.rawPath, tt.platform))
+		})
+	}
+}
+
+// ──────────────────────────────────────────────────────────
+// responsesSubpathSuffix
+// ──────────────────────────────────────────────────────────
+
+func TestResponsesSubpathSuffix(t *testing.T) {
+	tests := []struct {
+		raw  string
+		want string
+	}{
+		{"/v1/responses", ""},
+		{"/v1/responses/", ""},
+		{"/v1/responses/compact", "/compact"},
+		{"/openai/v1/responses/compact/detail", "/compact/detail"},
+		{"/v1/messages", ""},
+		{"", ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.raw, func(t *testing.T) {
+			require.Equal(t, tt.want, responsesSubpathSuffix(tt.raw))
+		})
+	}
+}
+
+// ──────────────────────────────────────────────────────────
+// InboundEndpointMiddleware + context helpers
+// ──────────────────────────────────────────────────────────
+
+func TestInboundEndpointMiddleware(t *testing.T) {
+	router := gin.New()
+	router.Use(InboundEndpointMiddleware())
+
+	var captured string
+	router.POST("/v1/messages", func(c *gin.Context) {
+		captured = GetInboundEndpoint(c)
+		c.Status(http.StatusOK)
+	})
+
+	req := httptest.NewRequest(http.MethodPost, "/v1/messages", nil)
+	rec := httptest.NewRecorder()
+	router.ServeHTTP(rec, req)
+
+	require.Equal(t, EndpointMessages, captured)
+}
+
+func TestGetInboundEndpoint_FallbackWithoutMiddleware(t *testing.T) {
+	rec := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(rec)
+	c.Request = httptest.NewRequest(http.MethodPost, "/antigravity/v1/messages", nil)
+
+	// Middleware did not run — fallback to normalizing c.Request.URL.Path.
+	got := GetInboundEndpoint(c)
+	require.Equal(t, EndpointMessages, got)
+}
+
+func TestGetUpstreamEndpoint_FullFlow(t *testing.T) {
+	rec := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(rec)
+	c.Request = httptest.NewRequest(http.MethodPost, "/openai/v1/responses/compact", nil)
+
+	// Simulate middleware.
+	c.Set(ctxKeyInboundEndpoint, NormalizeInboundEndpoint(c.Request.URL.Path))
+
+	got := GetUpstreamEndpoint(c, service.PlatformOpenAI)
+	require.Equal(t, "/v1/responses/compact", got)
+}

backend/internal/handler/gateway_handler.go

@@ -391,6 +391,8 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 			if fs.SwitchCount > 0 {
 				requestCtx = service.WithAccountSwitchCount(requestCtx, fs.SwitchCount, h.metadataBridgeEnabled())
 			}
+			// 记录 Forward 前已写入字节数，Forward 后若增加则说明 SSE 内容已发，禁止 failover
+			writerSizeBeforeForward := c.Writer.Size()
 			if account.Platform == service.PlatformAntigravity {
 				result, err = h.antigravityGatewayService.ForwardGemini(requestCtx, c, account, reqModel, "generateContent", reqStream, body, hasBoundSession)
 			} else {
@@ -402,6 +404,11 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 			if err != nil {
 				var failoverErr *service.UpstreamFailoverError
 				if errors.As(err, &failoverErr) {
+					// 流式内容已写入客户端，无法撤销，禁止 failover 以防止流拼接腐化
+					if c.Writer.Size() != writerSizeBeforeForward {
+						h.handleFailoverExhausted(c, failoverErr, service.PlatformGemini, true)
+						return
+					}
 					action := fs.HandleFailoverError(c.Request.Context(), h.gatewayService, account.ID, account.Platform, failoverErr)
 					switch action {
 					case FailoverContinue:
@@ -434,19 +441,29 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 			// 捕获请求信息（用于异步记录，避免在 goroutine 中访问 gin.Context）
 			userAgent := c.GetHeader("User-Agent")
 			clientIP := ip.GetClientIP(c)
+			requestPayloadHash := service.HashUsageRequestPayload(body)
+			inboundEndpoint := GetInboundEndpoint(c)
+			upstreamEndpoint := GetUpstreamEndpoint(c, account.Platform)
+
+			if result.ReasoningEffort == nil {
+				result.ReasoningEffort = service.NormalizeClaudeOutputEffort(parsedReq.OutputEffort)
+			}
 
 			// 使用量记录通过有界 worker 池提交，避免请求热路径创建无界 goroutine。
 			h.submitUsageRecordTask(func(ctx context.Context) {
 				if err := h.gatewayService.RecordUsage(ctx, &service.RecordUsageInput{
-					Result:            result,
-					APIKey:            apiKey,
-					User:              apiKey.User,
-					Account:           account,
-					Subscription:      subscription,
-					UserAgent:         userAgent,
-					IPAddress:         clientIP,
-					ForceCacheBilling: fs.ForceCacheBilling,
-					APIKeyService:     h.apiKeyService,
+					Result:             result,
+					APIKey:             apiKey,
+					User:               apiKey.User,
+					Account:            account,
+					Subscription:       subscription,
+					InboundEndpoint:    inboundEndpoint,
+					UpstreamEndpoint:   upstreamEndpoint,
+					UserAgent:          userAgent,
+					IPAddress:          clientIP,
+					RequestPayloadHash: requestPayloadHash,
+					ForceCacheBilling:  fs.ForceCacheBilling,
+					APIKeyService:      h.apiKeyService,
 				}); err != nil {
 					logger.L().With(
 						zap.String("component", "handler.gateway.messages"),
@@ -635,6 +652,8 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 			if fs.SwitchCount > 0 {
 				requestCtx = service.WithAccountSwitchCount(requestCtx, fs.SwitchCount, h.metadataBridgeEnabled())
 			}
+			// 记录 Forward 前已写入字节数，Forward 后若增加则说明 SSE 内容已发，禁止 failover
+			writerSizeBeforeForward := c.Writer.Size()
 			if account.Platform == service.PlatformAntigravity && account.Type != service.AccountTypeAPIKey {
 				result, err = h.antigravityGatewayService.Forward(requestCtx, c, account, body, hasBoundSession)
 			} else {
@@ -704,6 +723,11 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 				}
 				var failoverErr *service.UpstreamFailoverError
 				if errors.As(err, &failoverErr) {
+					// 流式内容已写入客户端，无法撤销，禁止 failover 以防止流拼接腐化
+					if c.Writer.Size() != writerSizeBeforeForward {
+						h.handleFailoverExhausted(c, failoverErr, account.Platform, true)
+						return
+					}
 					action := fs.HandleFailoverError(c.Request.Context(), h.gatewayService, account.ID, account.Platform, failoverErr)
 					switch action {
 					case FailoverContinue:
@@ -736,19 +760,29 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
 			// 捕获请求信息（用于异步记录，避免在 goroutine 中访问 gin.Context）
 			userAgent := c.GetHeader("User-Agent")
 			clientIP := ip.GetClientIP(c)
+			requestPayloadHash := service.HashUsageRequestPayload(body)
+			inboundEndpoint := GetInboundEndpoint(c)
+			upstreamEndpoint := GetUpstreamEndpoint(c, account.Platform)
+
+			if result.ReasoningEffort == nil {
+				result.ReasoningEffort = service.NormalizeClaudeOutputEffort(parsedReq.OutputEffort)
+			}
 
 			// 使用量记录通过有界 worker 池提交，避免请求热路径创建无界 goroutine。
 			h.submitUsageRecordTask(func(ctx context.Context) {
 				if err := h.gatewayService.RecordUsage(ctx, &service.RecordUsageInput{
-					Result:            result,
-					APIKey:            currentAPIKey,
-					User:              currentAPIKey.User,
-					Account:           account,
-					Subscription:      currentSubscription,
-					UserAgent:         userAgent,
-					IPAddress:         clientIP,
-					ForceCacheBilling: fs.ForceCacheBilling,
-					APIKeyService:     h.apiKeyService,
+					Result:             result,
+					APIKey:             currentAPIKey,
+					User:               currentAPIKey.User,
+					Account:            account,
+					Subscription:       currentSubscription,
+					InboundEndpoint:    inboundEndpoint,
+					UpstreamEndpoint:   upstreamEndpoint,
+					UserAgent:          userAgent,
+					IPAddress:          clientIP,
+					RequestPayloadHash: requestPayloadHash,
+					ForceCacheBilling:  fs.ForceCacheBilling,
+					APIKeyService:      h.apiKeyService,
 				}); err != nil {
 					logger.L().With(
 						zap.String("component", "handler.gateway.messages"),
@@ -909,7 +943,7 @@ func (h *GatewayHandler) parseUsageDateRange(c *gin.Context) (time.Time, time.Ti
 	}
 	if s := c.Query("end_date"); s != "" {
 		if t, err := timezone.ParseInLocation("2006-01-02", s); err == nil {
-			endTime = t.Add(24*time.Hour - time.Second) // end of day
+			endTime = t.AddDate(0, 0, 1) // half-open range upper bound
 		}
 	}
 	return startTime, endTime

backend/internal/handler/gateway_handler_stream_failover_test.go

@@ -0,0 +1,122 @@
+package handler
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// partialMessageStartSSE 模拟 handleStreamingResponse 已写入的首批 SSE 事件。
+const partialMessageStartSSE = "event: message_start\ndata: {\"type\":\"message_start\",\"message\":{\"id\":\"msg_01\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[],\"model\":\"claude-sonnet-4-5\",\"stop_reason\":null,\"stop_sequence\":null,\"usage\":{\"input_tokens\":10,\"output_tokens\":1}}}\n\n" +
+	"event: content_block_start\ndata: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"text\",\"text\":\"\"}}\n\n"
+
+// TestStreamWrittenGuard_MessagesPath_AbortFailoverOnSSEContentWritten 验证：
+// 当 Forward 在返回 UpstreamFailoverError 前已向客户端写入 SSE 内容时，
+// 故障转移保护逻辑必须终止循环并发送 SSE 错误事件，而不是进行下一次 Forward。
+// 具体验证：
+//  1. c.Writer.Size() 检测条件正确触发（字节数已增加）
+//  2. handleFailoverExhausted 以 streamStarted=true 调用后，响应体以 SSE 错误事件结尾
+//  3. 响应体中只出现一个 message_start，不存在第二个（防止流拼接腐化）
+func TestStreamWrittenGuard_MessagesPath_AbortFailoverOnSSEContentWritten(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	w := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(w)
+	c.Request = httptest.NewRequest(http.MethodPost, "/v1/messages", nil)
+
+	// 步骤 1：记录 Forward 前的 writer size（模拟 writerSizeBeforeForward := c.Writer.Size()）
+	sizeBeforeForward := c.Writer.Size()
+	require.Equal(t, -1, sizeBeforeForward, "gin writer 初始 Size 应为 -1（未写入任何字节）")
+
+	// 步骤 2：模拟 Forward 已向客户端写入部分 SSE 内容（message_start + content_block_start）
+	_, err := c.Writer.Write([]byte(partialMessageStartSSE))
+	require.NoError(t, err)
+
+	// 步骤 3：验证守卫条件成立（c.Writer.Size() != sizeBeforeForward）
+	require.NotEqual(t, sizeBeforeForward, c.Writer.Size(),
+		"写入 SSE 内容后 writer size 必须增加，守卫条件应为 true")
+
+	// 步骤 4：模拟 UpstreamFailoverError（上游在流中途返回 403）
+	failoverErr := &service.UpstreamFailoverError{
+		StatusCode:   http.StatusForbidden,
+		ResponseBody: []byte(`{"error":{"type":"permission_error","message":"forbidden"}}`),
+	}
+
+	// 步骤 5：守卫触发 → 调用 handleFailoverExhausted，streamStarted=true
+	h := &GatewayHandler{}
+	h.handleFailoverExhausted(c, failoverErr, service.PlatformAnthropic, true)
+
+	body := w.Body.String()
+
+	// 断言 A：响应体中包含最初写入的 message_start SSE 事件行
+	require.Contains(t, body, "event: message_start", "响应体应包含已写入的 message_start SSE 事件")
+
+	// 断言 B：响应体以 SSE 错误事件结尾（data: {"type":"error",...}\n\n）
+	require.True(t, strings.HasSuffix(strings.TrimRight(body, "\n"), "}"),
+		"响应体应以 JSON 对象结尾（SSE error event 的 data 字段）")
+	require.Contains(t, body, `"type":"error"`, "响应体末尾必须包含 SSE 错误事件")
+
+	// 断言 C：SSE event 行 "event: message_start" 只出现一次（防止双 message_start 拼接腐化）
+	firstIdx := strings.Index(body, "event: message_start")
+	lastIdx := strings.LastIndex(body, "event: message_start")
+	assert.Equal(t, firstIdx, lastIdx,
+		"响应体中 'event: message_start' 必须只出现一次，不得因 failover 拼接导致两次")
+}
+
+// TestStreamWrittenGuard_GeminiPath_AbortFailoverOnSSEContentWritten 与上述测试相同，
+// 验证 Gemini 路径使用 service.PlatformGemini（而非 account.Platform）时行为一致。
+func TestStreamWrittenGuard_GeminiPath_AbortFailoverOnSSEContentWritten(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	w := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(w)
+	c.Request = httptest.NewRequest(http.MethodPost, "/v1beta/models/gemini-2.0-flash:streamGenerateContent", nil)
+
+	sizeBeforeForward := c.Writer.Size()
+
+	_, err := c.Writer.Write([]byte(partialMessageStartSSE))
+	require.NoError(t, err)
+
+	require.NotEqual(t, sizeBeforeForward, c.Writer.Size())
+
+	failoverErr := &service.UpstreamFailoverError{
+		StatusCode: http.StatusForbidden,
+	}
+
+	h := &GatewayHandler{}
+	h.handleFailoverExhausted(c, failoverErr, service.PlatformGemini, true)
+
+	body := w.Body.String()
+
+	require.Contains(t, body, "event: message_start")
+	require.Contains(t, body, `"type":"error"`)
+
+	firstIdx := strings.Index(body, "event: message_start")
+	lastIdx := strings.LastIndex(body, "event: message_start")
+	assert.Equal(t, firstIdx, lastIdx, "Gemini 路径不得出现双 message_start")
+}
+
+// TestStreamWrittenGuard_NoByteWritten_GuardNotTriggered 验证反向场景：
+// 当 Forward 返回 UpstreamFailoverError 时若未向客户端写入任何 SSE 内容，
+// 守卫条件（c.Writer.Size() != sizeBeforeForward）为 false，不应中止 failover。
+func TestStreamWrittenGuard_NoByteWritten_GuardNotTriggered(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	w := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(w)
+	c.Request = httptest.NewRequest(http.MethodPost, "/v1/messages", nil)
+
+	// 模拟 writerSizeBeforeForward：初始为 -1
+	sizeBeforeForward := c.Writer.Size()
+
+	// Forward 未写入任何字节直接返回错误（例如 401 发生在连接建立前）
+	// c.Writer.Size() 仍为 -1
+
+	// 守卫条件：sizeBeforeForward == c.Writer.Size() → 不触发
+	guardTriggered := c.Writer.Size() != sizeBeforeForward
+	require.False(t, guardTriggered,
+		"未写入任何字节时，守卫条件必须为 false，应允许正常 failover 继续")
+}

backend/internal/handler/gateway_handler_warmup_intercept_unit_test.go

@@ -76,7 +76,7 @@ func (f *fakeGroupRepo) ListActiveByPlatform(context.Context, string) ([]service
 	return nil, nil
 }
 func (f *fakeGroupRepo) ExistsByName(context.Context, string) (bool, error)    { return false, nil }
-func (f *fakeGroupRepo) GetAccountCount(context.Context, int64) (int64, error) { return 0, nil }
+func (f *fakeGroupRepo) GetAccountCount(context.Context, int64) (int64, int64, error) { return 0, 0, nil }
 func (f *fakeGroupRepo) DeleteAccountGroupsByGroupID(context.Context, int64) (int64, error) {
 	return 0, nil
 }
@@ -139,6 +139,7 @@ func newTestGatewayHandler(t *testing.T, group *service.Group, accounts []*servi
 		nil, // accountRepo (not used: scheduler snapshot hit)
 		&fakeGroupRepo{group: group},
 		nil, // usageLogRepo
+		nil, // usageBillingRepo
 		nil, // userRepo
 		nil, // userSubRepo
 		nil, // userGroupRateRepo

backend/internal/handler/gateway_helper_hotpath_test.go

@@ -136,7 +136,7 @@ func validClaudeCodeBodyJSON() []byte {
 	return []byte(`{
 		"model":"claude-3-5-sonnet-20241022",
 		"system":[{"text":"You are Claude Code, Anthropic's official CLI for Claude."}],
-		"metadata":{"user_id":"user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_abc-123"}
+		"metadata":{"user_id":"user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"}
 	}`)
 }
 
@@ -190,7 +190,7 @@ func TestSetClaudeCodeClientContext_ReuseParsedRequestAndContextCache(t *testing
 			System: []any{
 				map[string]any{"text": "You are Claude Code, Anthropic's official CLI for Claude."},
 			},
-			MetadataUserID: "user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_abc-123",
+			MetadataUserID: "user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
 		}
 
 		// body 非法 JSON，如果函数复用 parsedReq 成功则仍应判定为 Claude Code。
@@ -209,7 +209,7 @@ func TestSetClaudeCodeClientContext_ReuseParsedRequestAndContextCache(t *testing
 			"system": []any{
 				map[string]any{"text": "You are Claude Code, Anthropic's official CLI for Claude."},
 			},
-			"metadata": map[string]any{"user_id": "user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_abc-123"},
+			"metadata": map[string]any{"user_id": "user_aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa_account__session_aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"},
 		})
 
 		SetClaudeCodeClientContext(c, []byte(`{invalid`), nil)

backend/internal/handler/gemini_v1beta_handler.go

@@ -503,6 +503,9 @@ func (h *GatewayHandler) GeminiV1BetaModels(c *gin.Context) {
 		}
 
 		// 使用量记录通过有界 worker 池提交，避免请求热路径创建无界 goroutine。
+		requestPayloadHash := service.HashUsageRequestPayload(body)
+		inboundEndpoint := GetInboundEndpoint(c)
+		upstreamEndpoint := GetUpstreamEndpoint(c, account.Platform)
 		h.submitUsageRecordTask(func(ctx context.Context) {
 			if err := h.gatewayService.RecordUsageWithLongContext(ctx, &service.RecordUsageLongContextInput{
 				Result:                result,
@@ -510,8 +513,11 @@ func (h *GatewayHandler) GeminiV1BetaModels(c *gin.Context) {
 				User:                  apiKey.User,
 				Account:               account,
 				Subscription:          subscription,
+				InboundEndpoint:       inboundEndpoint,
+				UpstreamEndpoint:      upstreamEndpoint,
 				UserAgent:             userAgent,
 				IPAddress:             clientIP,
+				RequestPayloadHash:    requestPayloadHash,
 				LongContextThreshold:  200000, // Gemini 200K 阈值
 				LongContextMultiplier: 2.0,    // 超出部分双倍计费
 				ForceCacheBilling:     fs.ForceCacheBilling,

backend/internal/handler/handler.go

@@ -12,6 +12,7 @@ type AdminHandlers struct {
 	Account          *admin.AccountHandler
 	Announcement     *admin.AnnouncementHandler
 	DataManagement   *admin.DataManagementHandler
+	Backup           *admin.BackupHandler
 	OAuth            *admin.OAuthHandler
 	OpenAIOAuth      *admin.OpenAIOAuthHandler
 	GeminiOAuth      *admin.GeminiOAuthHandler

backend/internal/handler/openai_chat_completions.go

@@ -181,13 +181,7 @@ func (h *OpenAIGatewayHandler) ChatCompletions(c *gin.Context) {
 		service.SetOpsLatencyMs(c, service.OpsRoutingLatencyMsKey, time.Since(routingStart).Milliseconds())
 		forwardStart := time.Now()
 
-		defaultMappedModel := ""
-		if apiKey.Group != nil {
-			defaultMappedModel = apiKey.Group.DefaultMappedModel
-		}
-		if fallbackModel := c.GetString("openai_chat_completions_fallback_model"); fallbackModel != "" {
-			defaultMappedModel = fallbackModel
-		}
+		defaultMappedModel := c.GetString("openai_chat_completions_fallback_model")
 		result, err := h.gatewayService.ForwardAsChatCompletions(c.Request.Context(), c, account, body, promptCacheKey, defaultMappedModel)
 
 		forwardDurationMs := time.Since(forwardStart).Milliseconds()
@@ -262,14 +256,16 @@ func (h *OpenAIGatewayHandler) ChatCompletions(c *gin.Context) {
 
 		h.submitUsageRecordTask(func(ctx context.Context) {
 			if err := h.gatewayService.RecordUsage(ctx, &service.OpenAIRecordUsageInput{
-				Result:        result,
-				APIKey:        apiKey,
-				User:          apiKey.User,
-				Account:       account,
-				Subscription:  subscription,
-				UserAgent:     userAgent,
-				IPAddress:     clientIP,
-				APIKeyService: h.apiKeyService,
+				Result:           result,
+				APIKey:           apiKey,
+				User:             apiKey.User,
+				Account:          account,
+				Subscription:     subscription,
+				InboundEndpoint:  GetInboundEndpoint(c),
+				UpstreamEndpoint: GetUpstreamEndpoint(c, account.Platform),
+				UserAgent:        userAgent,
+				IPAddress:        clientIP,
+				APIKeyService:    h.apiKeyService,
 			}); err != nil {
 				logger.L().With(
 					zap.String("component", "handler.openai_gateway.chat_completions"),

backend/internal/handler/openai_gateway_endpoint_normalization_test.go

@@ -0,0 +1,56 @@
+package handler
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/require"
+)
+
+// TestOpenAIUpstreamEndpoint_ViaGetUpstreamEndpoint verifies that the
+// unified GetUpstreamEndpoint helper produces the same results as the
+// former normalizedOpenAIUpstreamEndpoint for OpenAI platform requests.
+func TestOpenAIUpstreamEndpoint_ViaGetUpstreamEndpoint(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+
+	tests := []struct {
+		name string
+		path string
+		want string
+	}{
+		{
+			name: "responses root maps to responses upstream",
+			path: "/v1/responses",
+			want: EndpointResponses,
+		},
+		{
+			name: "responses compact keeps compact suffix",
+			path: "/openai/v1/responses/compact",
+			want: "/v1/responses/compact",
+		},
+		{
+			name: "responses nested suffix preserved",
+			path: "/openai/v1/responses/compact/detail",
+			want: "/v1/responses/compact/detail",
+		},
+		{
+			name: "non responses path uses platform fallback",
+			path: "/v1/messages",
+			want: EndpointResponses,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			rec := httptest.NewRecorder()
+			c, _ := gin.CreateTestContext(rec)
+			c.Request = httptest.NewRequest(http.MethodPost, tt.path, nil)
+
+			got := GetUpstreamEndpoint(c, service.PlatformOpenAI)
+			require.Equal(t, tt.want, got)
+		})
+	}
+}

backend/internal/handler/openai_gateway_handler.go

@@ -352,18 +352,22 @@ func (h *OpenAIGatewayHandler) Responses(c *gin.Context) {
 		// 捕获请求信息（用于异步记录，避免在 goroutine 中访问 gin.Context）
 		userAgent := c.GetHeader("User-Agent")
 		clientIP := ip.GetClientIP(c)
+		requestPayloadHash := service.HashUsageRequestPayload(body)
 
 		// 使用量记录通过有界 worker 池提交，避免请求热路径创建无界 goroutine。
 		h.submitUsageRecordTask(func(ctx context.Context) {
 			if err := h.gatewayService.RecordUsage(ctx, &service.OpenAIRecordUsageInput{
-				Result:        result,
-				APIKey:        apiKey,
-				User:          apiKey.User,
-				Account:       account,
-				Subscription:  subscription,
-				UserAgent:     userAgent,
-				IPAddress:     clientIP,
-				APIKeyService: h.apiKeyService,
+				Result:             result,
+				APIKey:             apiKey,
+				User:               apiKey.User,
+				Account:            account,
+				Subscription:       subscription,
+				InboundEndpoint:    GetInboundEndpoint(c),
+				UpstreamEndpoint:   GetUpstreamEndpoint(c, account.Platform),
+				UserAgent:          userAgent,
+				IPAddress:          clientIP,
+				RequestPayloadHash: requestPayloadHash,
+				APIKeyService:      h.apiKeyService,
 			}); err != nil {
 				logger.L().With(
 					zap.String("component", "handler.openai_gateway.responses"),
@@ -653,14 +657,9 @@ func (h *OpenAIGatewayHandler) Messages(c *gin.Context) {
 		service.SetOpsLatencyMs(c, service.OpsRoutingLatencyMsKey, time.Since(routingStart).Milliseconds())
 		forwardStart := time.Now()
 
-		defaultMappedModel := ""
-		if apiKey.Group != nil {
-			defaultMappedModel = apiKey.Group.DefaultMappedModel
-		}
-		// 如果使用了降级模型调度，强制使用降级模型
-		if fallbackModel := c.GetString("openai_messages_fallback_model"); fallbackModel != "" {
-			defaultMappedModel = fallbackModel
-		}
+		// 仅在调度时实际触发了降级（原模型无可用账号、改用默认模型重试成功）时，
+		// 才将降级模型传给 Forward 层做模型替换；否则保持用户请求的原始模型。
+		defaultMappedModel := c.GetString("openai_messages_fallback_model")
 		result, err := h.gatewayService.ForwardAsAnthropic(c.Request.Context(), c, account, body, promptCacheKey, defaultMappedModel)
 
 		forwardDurationMs := time.Since(forwardStart).Milliseconds()
@@ -732,17 +731,21 @@ func (h *OpenAIGatewayHandler) Messages(c *gin.Context) {
 
 		userAgent := c.GetHeader("User-Agent")
 		clientIP := ip.GetClientIP(c)
+		requestPayloadHash := service.HashUsageRequestPayload(body)
 
 		h.submitUsageRecordTask(func(ctx context.Context) {
 			if err := h.gatewayService.RecordUsage(ctx, &service.OpenAIRecordUsageInput{
-				Result:        result,
-				APIKey:        apiKey,
-				User:          apiKey.User,
-				Account:       account,
-				Subscription:  subscription,
-				UserAgent:     userAgent,
-				IPAddress:     clientIP,
-				APIKeyService: h.apiKeyService,
+				Result:             result,
+				APIKey:             apiKey,
+				User:               apiKey.User,
+				Account:            account,
+				Subscription:       subscription,
+				InboundEndpoint:    GetInboundEndpoint(c),
+				UpstreamEndpoint:   GetUpstreamEndpoint(c, account.Platform),
+				UserAgent:          userAgent,
+				IPAddress:          clientIP,
+				RequestPayloadHash: requestPayloadHash,
+				APIKeyService:      h.apiKeyService,
 			}); err != nil {
 				logger.L().With(
 					zap.String("component", "handler.openai_gateway.messages"),
@@ -1231,14 +1234,17 @@ func (h *OpenAIGatewayHandler) ResponsesWebSocket(c *gin.Context) {
 			h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, result.FirstTokenMs)
 			h.submitUsageRecordTask(func(taskCtx context.Context) {
 				if err := h.gatewayService.RecordUsage(taskCtx, &service.OpenAIRecordUsageInput{
-					Result:        result,
-					APIKey:        apiKey,
-					User:          apiKey.User,
-					Account:       account,
-					Subscription:  subscription,
-					UserAgent:     userAgent,
-					IPAddress:     clientIP,
-					APIKeyService: h.apiKeyService,
+					Result:             result,
+					APIKey:             apiKey,
+					User:               apiKey.User,
+					Account:            account,
+					Subscription:       subscription,
+					InboundEndpoint:    GetInboundEndpoint(c),
+					UpstreamEndpoint:   GetUpstreamEndpoint(c, account.Platform),
+					UserAgent:          userAgent,
+					IPAddress:          clientIP,
+					RequestPayloadHash: service.HashUsageRequestPayload(firstMessage),
+					APIKeyService:      h.apiKeyService,
 				}); err != nil {
 					reqLog.Error("openai.websocket_record_usage_failed",
 						zap.Int64("account_id", account.ID),

backend/internal/handler/ops_error_logger.go

@@ -26,6 +26,22 @@ const (
 	opsStreamKey      = "ops_stream"
 	opsRequestBodyKey = "ops_request_body"
 	opsAccountIDKey   = "ops_account_id"
+
+	// 错误过滤匹配常量 — shouldSkipOpsErrorLog 和错误分类共用
+	opsErrContextCanceled            = "context canceled"
+	opsErrNoAvailableAccounts        = "no available accounts"
+	opsErrInvalidAPIKey              = "invalid_api_key"
+	opsErrAPIKeyRequired             = "api_key_required"
+	opsErrInsufficientBalance        = "insufficient balance"
+	opsErrInsufficientAccountBalance = "insufficient account balance"
+	opsErrInsufficientQuota          = "insufficient_quota"
+
+	// 上游错误码常量 — 错误分类 (normalizeOpsErrorType / classifyOpsPhase / classifyOpsIsBusinessLimited)
+	opsCodeInsufficientBalance  = "INSUFFICIENT_BALANCE"
+	opsCodeUsageLimitExceeded   = "USAGE_LIMIT_EXCEEDED"
+	opsCodeSubscriptionNotFound = "SUBSCRIPTION_NOT_FOUND"
+	opsCodeSubscriptionInvalid  = "SUBSCRIPTION_INVALID"
+	opsCodeUserInactive         = "USER_INACTIVE"
 )
 
 const (
@@ -1024,9 +1040,9 @@ func normalizeOpsErrorType(errType string, code string) string {
 		return errType
 	}
 	switch strings.TrimSpace(code) {
-	case "INSUFFICIENT_BALANCE":
+	case opsCodeInsufficientBalance:
 		return "billing_error"
-	case "USAGE_LIMIT_EXCEEDED", "SUBSCRIPTION_NOT_FOUND", "SUBSCRIPTION_INVALID":
+	case opsCodeUsageLimitExceeded, opsCodeSubscriptionNotFound, opsCodeSubscriptionInvalid:
 		return "subscription_error"
 	default:
 		return "api_error"
@@ -1038,7 +1054,7 @@ func classifyOpsPhase(errType, message, code string) string {
 	// Standardized phases: request|auth|routing|upstream|network|internal
 	// Map billing/concurrency/response => request; scheduling => routing.
 	switch strings.TrimSpace(code) {
-	case "INSUFFICIENT_BALANCE", "USAGE_LIMIT_EXCEEDED", "SUBSCRIPTION_NOT_FOUND", "SUBSCRIPTION_INVALID":
+	case opsCodeInsufficientBalance, opsCodeUsageLimitExceeded, opsCodeSubscriptionNotFound, opsCodeSubscriptionInvalid:
 		return "request"
 	}
 
@@ -1057,7 +1073,7 @@ func classifyOpsPhase(errType, message, code string) string {
 	case "upstream_error", "overloaded_error":
 		return "upstream"
 	case "api_error":
-		if strings.Contains(msg, "no available accounts") {
+		if strings.Contains(msg, opsErrNoAvailableAccounts) {
 			return "routing"
 		}
 		return "internal"
@@ -1103,7 +1119,7 @@ func classifyOpsIsRetryable(errType string, statusCode int) bool {
 
 func classifyOpsIsBusinessLimited(errType, phase, code string, status int, message string) bool {
 	switch strings.TrimSpace(code) {
-	case "INSUFFICIENT_BALANCE", "USAGE_LIMIT_EXCEEDED", "SUBSCRIPTION_NOT_FOUND", "SUBSCRIPTION_INVALID", "USER_INACTIVE":
+	case opsCodeInsufficientBalance, opsCodeUsageLimitExceeded, opsCodeSubscriptionNotFound, opsCodeSubscriptionInvalid, opsCodeUserInactive:
 		return true
 	}
 	if phase == "billing" || phase == "concurrency" {
@@ -1197,21 +1213,30 @@ func shouldSkipOpsErrorLog(ctx context.Context, ops *service.OpsService, message
 
 	// Check if context canceled errors should be ignored (client disconnects)
 	if settings.IgnoreContextCanceled {
-		if strings.Contains(msgLower, "context canceled") || strings.Contains(bodyLower, "context canceled") {
+		if strings.Contains(msgLower, opsErrContextCanceled) || strings.Contains(bodyLower, opsErrContextCanceled) {
 			return true
 		}
 	}
 
 	// Check if "no available accounts" errors should be ignored
 	if settings.IgnoreNoAvailableAccounts {
-		if strings.Contains(msgLower, "no available accounts") || strings.Contains(bodyLower, "no available accounts") {
+		if strings.Contains(msgLower, opsErrNoAvailableAccounts) || strings.Contains(bodyLower, opsErrNoAvailableAccounts) {
 			return true
 		}
 	}
 
 	// Check if invalid/missing API key errors should be ignored (user misconfiguration)
 	if settings.IgnoreInvalidApiKeyErrors {
-		if strings.Contains(bodyLower, "invalid_api_key") || strings.Contains(bodyLower, "api_key_required") {
+		if strings.Contains(bodyLower, opsErrInvalidAPIKey) || strings.Contains(bodyLower, opsErrAPIKeyRequired) {
+			return true
+		}
+	}
+
+	// Check if insufficient balance errors should be ignored
+	if settings.IgnoreInsufficientBalanceErrors {
+		if strings.Contains(bodyLower, opsErrInsufficientBalance) || strings.Contains(bodyLower, opsErrInsufficientAccountBalance) ||
+			strings.Contains(bodyLower, opsErrInsufficientQuota) ||
+			strings.Contains(msgLower, opsErrInsufficientBalance) || strings.Contains(msgLower, opsErrInsufficientAccountBalance) {
 			return true
 		}
 	}

backend/internal/handler/setting_handler.go

@@ -54,6 +54,7 @@ func (h *SettingHandler) GetPublicSettings(c *gin.Context) {
 		CustomMenuItems:                  dto.ParseUserVisibleMenuItems(settings.CustomMenuItems),
 		LinuxDoOAuthEnabled:              settings.LinuxDoOAuthEnabled,
 		SoraClientEnabled:                settings.SoraClientEnabled,
+		BackendModeEnabled:               settings.BackendModeEnabled,
 		Version:                          h.version,
 	})
 }

backend/internal/handler/sora_client_handler_test.go

@@ -2206,7 +2206,7 @@ func (s *stubSoraClientForHandler) GetVideoTask(_ context.Context, _ *service.Ac
 // newMinimalGatewayService 创建仅包含 accountRepo 的最小 GatewayService（用于测试 SelectAccountForModel）。
 func newMinimalGatewayService(accountRepo service.AccountRepository) *service.GatewayService {
 	return service.NewGatewayService(
-		accountRepo, nil, nil, nil, nil, nil, nil, nil,
+		accountRepo, nil, nil, nil, nil, nil, nil, nil, nil,
 		nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil,
 	)
 }

backend/internal/handler/sora_gateway_handler.go

@@ -399,17 +399,23 @@ func (h *SoraGatewayHandler) ChatCompletions(c *gin.Context) {
 
 		userAgent := c.GetHeader("User-Agent")
 		clientIP := ip.GetClientIP(c)
+		requestPayloadHash := service.HashUsageRequestPayload(body)
+		inboundEndpoint := GetInboundEndpoint(c)
+		upstreamEndpoint := GetUpstreamEndpoint(c, account.Platform)
 
 		// 使用量记录通过有界 worker 池提交，避免请求热路径创建无界 goroutine。
 		h.submitUsageRecordTask(func(ctx context.Context) {
 			if err := h.gatewayService.RecordUsage(ctx, &service.RecordUsageInput{
-				Result:       result,
-				APIKey:       apiKey,
-				User:         apiKey.User,
-				Account:      account,
-				Subscription: subscription,
-				UserAgent:    userAgent,
-				IPAddress:    clientIP,
+				Result:             result,
+				APIKey:             apiKey,
+				User:               apiKey.User,
+				Account:            account,
+				Subscription:       subscription,
+				InboundEndpoint:    inboundEndpoint,
+				UpstreamEndpoint:   upstreamEndpoint,
+				UserAgent:          userAgent,
+				IPAddress:          clientIP,
+				RequestPayloadHash: requestPayloadHash,
 			}); err != nil {
 				logger.L().With(
 					zap.String("component", "handler.sora_gateway.chat_completions"),

backend/internal/handler/sora_gateway_handler_test.go

@@ -273,8 +273,8 @@ func (r *stubGroupRepo) ListActiveByPlatform(ctx context.Context, platform strin
 func (r *stubGroupRepo) ExistsByName(ctx context.Context, name string) (bool, error) {
 	return false, nil
 }
-func (r *stubGroupRepo) GetAccountCount(ctx context.Context, groupID int64) (int64, error) {
-	return 0, nil
+func (r *stubGroupRepo) GetAccountCount(ctx context.Context, groupID int64) (int64, int64, error) {
+	return 0, 0, nil
 }
 func (r *stubGroupRepo) DeleteAccountGroupsByGroupID(ctx context.Context, groupID int64) (int64, error) {
 	return 0, nil
@@ -334,15 +334,32 @@ func (s *stubUsageLogRepo) GetUsageTrendWithFilters(ctx context.Context, startTi
 func (s *stubUsageLogRepo) GetModelStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, requestType *int16, stream *bool, billingType *int8) ([]usagestats.ModelStat, error) {
 	return nil, nil
 }
+
+func (s *stubUsageLogRepo) GetEndpointStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, model string, requestType *int16, stream *bool, billingType *int8) ([]usagestats.EndpointStat, error) {
+	return []usagestats.EndpointStat{}, nil
+}
+
+func (s *stubUsageLogRepo) GetUpstreamEndpointStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, model string, requestType *int16, stream *bool, billingType *int8) ([]usagestats.EndpointStat, error) {
+	return []usagestats.EndpointStat{}, nil
+}
 func (s *stubUsageLogRepo) GetGroupStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, requestType *int16, stream *bool, billingType *int8) ([]usagestats.GroupStat, error) {
 	return nil, nil
 }
+func (s *stubUsageLogRepo) GetUserBreakdownStats(ctx context.Context, startTime, endTime time.Time, dim usagestats.UserBreakdownDimension, limit int) ([]usagestats.UserBreakdownItem, error) {
+	return nil, nil
+}
+func (s *stubUsageLogRepo) GetAllGroupUsageSummary(ctx context.Context, todayStart time.Time) ([]usagestats.GroupUsageSummary, error) {
+	return nil, nil
+}
 func (s *stubUsageLogRepo) GetAPIKeyUsageTrend(ctx context.Context, startTime, endTime time.Time, granularity string, limit int) ([]usagestats.APIKeyUsageTrendPoint, error) {
 	return nil, nil
 }
 func (s *stubUsageLogRepo) GetUserUsageTrend(ctx context.Context, startTime, endTime time.Time, granularity string, limit int) ([]usagestats.UserUsageTrendPoint, error) {
 	return nil, nil
 }
+func (s *stubUsageLogRepo) GetUserSpendingRanking(ctx context.Context, startTime, endTime time.Time, limit int) (*usagestats.UserSpendingRankingResponse, error) {
+	return nil, nil
+}
 func (s *stubUsageLogRepo) GetBatchUserUsageStats(ctx context.Context, userIDs []int64, startTime, endTime time.Time) (map[int64]*usagestats.BatchUserUsageStats, error) {
 	return nil, nil
 }
@@ -431,6 +448,7 @@ func TestSoraGatewayHandler_ChatCompletions(t *testing.T) {
 		nil,
 		nil,
 		nil,
+		nil,
 		testutil.StubGatewayCache{},
 		cfg,
 		nil,

backend/internal/handler/usage_handler.go

@@ -114,8 +114,8 @@ func (h *UsageHandler) List(c *gin.Context) {
 			response.BadRequest(c, "Invalid end_date format, use YYYY-MM-DD")
 			return
 		}
-		// Set end time to end of day
-		t = t.Add(24*time.Hour - time.Nanosecond)
+		// Use half-open range [start, end), move to next calendar day start (DST-safe).
+		t = t.AddDate(0, 0, 1)
 		endTime = &t
 	}
 
@@ -227,8 +227,8 @@ func (h *UsageHandler) Stats(c *gin.Context) {
 			response.BadRequest(c, "Invalid end_date format, use YYYY-MM-DD")
 			return
 		}
-		// 设置结束时间为当天结束
-		endTime = endTime.Add(24*time.Hour - time.Nanosecond)
+		// 与 SQL 条件 created_at < end 对齐，使用次日 00:00 作为上边界（DST-safe）。
+		endTime = endTime.AddDate(0, 0, 1)
 	} else {
 		// 使用 period 参数
 		period := c.DefaultQuery("period", "today")

backend/internal/handler/wire.go

@@ -15,6 +15,7 @@ func ProvideAdminHandlers(
 	accountHandler *admin.AccountHandler,
 	announcementHandler *admin.AnnouncementHandler,
 	dataManagementHandler *admin.DataManagementHandler,
+	backupHandler *admin.BackupHandler,
 	oauthHandler *admin.OAuthHandler,
 	openaiOAuthHandler *admin.OpenAIOAuthHandler,
 	geminiOAuthHandler *admin.GeminiOAuthHandler,
@@ -39,6 +40,7 @@ func ProvideAdminHandlers(
 		Account:          accountHandler,
 		Announcement:     announcementHandler,
 		DataManagement:   dataManagementHandler,
+		Backup:           backupHandler,
 		OAuth:            oauthHandler,
 		OpenAIOAuth:      openaiOAuthHandler,
 		GeminiOAuth:      geminiOAuthHandler,
@@ -128,6 +130,7 @@ var ProviderSet = wire.NewSet(
 	admin.NewAccountHandler,
 	admin.NewAnnouncementHandler,
 	admin.NewDataManagementHandler,
+	admin.NewBackupHandler,
 	admin.NewOAuthHandler,
 	admin.NewOpenAIOAuthHandler,
 	admin.NewGeminiOAuthHandler,

backend/internal/pkg/antigravity/client.go

@@ -19,6 +19,16 @@ import (
 	"github.com/Wei-Shaw/sub2api/internal/pkg/proxyutil"
 )
 
+// ForbiddenError 表示上游返回 403 Forbidden
+type ForbiddenError struct {
+	StatusCode int
+	Body       string
+}
+
+func (e *ForbiddenError) Error() string {
+	return fmt.Sprintf("fetchAvailableModels 失败 (HTTP %d): %s", e.StatusCode, e.Body)
+}
+
 // NewAPIRequestWithURL 使用指定的 base URL 创建 Antigravity API 请求（v1internal 端点）
 func NewAPIRequestWithURL(ctx context.Context, baseURL, action, accessToken string, body []byte) (*http.Request, error) {
 	// 构建 URL，流式请求添加 ?alt=sse 参数
@@ -114,10 +124,68 @@ type IneligibleTier struct {
 type LoadCodeAssistResponse struct {
 	CloudAICompanionProject string            `json:"cloudaicompanionProject"`
 	CurrentTier             *TierInfo         `json:"currentTier,omitempty"`
-	PaidTier                *TierInfo         `json:"paidTier,omitempty"`
+	PaidTier                *PaidTierInfo     `json:"paidTier,omitempty"`
 	IneligibleTiers         []*IneligibleTier `json:"ineligibleTiers,omitempty"`
 }
 
+// PaidTierInfo 付费等级信息，包含 AI Credits 余额。
+type PaidTierInfo struct {
+	ID               string            `json:"id"`
+	Name             string            `json:"name"`
+	Description      string            `json:"description"`
+	AvailableCredits []AvailableCredit `json:"availableCredits,omitempty"`
+}
+
+// UnmarshalJSON 兼容 paidTier 既可能是字符串也可能是对象的情况。
+func (p *PaidTierInfo) UnmarshalJSON(data []byte) error {
+	data = bytes.TrimSpace(data)
+	if len(data) == 0 || string(data) == "null" {
+		return nil
+	}
+	if data[0] == '"' {
+		var id string
+		if err := json.Unmarshal(data, &id); err != nil {
+			return err
+		}
+		p.ID = id
+		return nil
+	}
+	type alias PaidTierInfo
+	var raw alias
+	if err := json.Unmarshal(data, &raw); err != nil {
+		return err
+	}
+	*p = PaidTierInfo(raw)
+	return nil
+}
+
+// AvailableCredit 表示一条 AI Credits 余额记录。
+type AvailableCredit struct {
+	CreditType                  string `json:"creditType,omitempty"`
+	CreditAmount                string `json:"creditAmount,omitempty"`
+	MinimumCreditAmountForUsage string `json:"minimumCreditAmountForUsage,omitempty"`
+}
+
+// GetAmount 将 creditAmount 解析为浮点数。
+func (c *AvailableCredit) GetAmount() float64 {
+	if c.CreditAmount == "" {
+		return 0
+	}
+	var value float64
+	_, _ = fmt.Sscanf(c.CreditAmount, "%f", &value)
+	return value
+}
+
+// GetMinimumAmount 将 minimumCreditAmountForUsage 解析为浮点数。
+func (c *AvailableCredit) GetMinimumAmount() float64 {
+	if c.MinimumCreditAmountForUsage == "" {
+		return 0
+	}
+	var value float64
+	_, _ = fmt.Sscanf(c.MinimumCreditAmountForUsage, "%f", &value)
+	return value
+}
+
 // OnboardUserRequest onboardUser 请求
 type OnboardUserRequest struct {
 	TierID   string `json:"tierId"`
@@ -147,6 +215,14 @@ func (r *LoadCodeAssistResponse) GetTier() string {
 	return ""
 }
 
+// GetAvailableCredits 返回 paid tier 中的 AI Credits 余额列表。
+func (r *LoadCodeAssistResponse) GetAvailableCredits() []AvailableCredit {
+	if r.PaidTier == nil {
+		return nil
+	}
+	return r.PaidTier.AvailableCredits
+}
+
 // Client Antigravity API 客户端
 type Client struct {
 	httpClient *http.Client
@@ -514,7 +590,20 @@ type ModelQuotaInfo struct {
 
 // ModelInfo 模型信息
 type ModelInfo struct {
-	QuotaInfo *ModelQuotaInfo `json:"quotaInfo,omitempty"`
+	QuotaInfo          *ModelQuotaInfo `json:"quotaInfo,omitempty"`
+	DisplayName        string          `json:"displayName,omitempty"`
+	SupportsImages     *bool           `json:"supportsImages,omitempty"`
+	SupportsThinking   *bool           `json:"supportsThinking,omitempty"`
+	ThinkingBudget     *int            `json:"thinkingBudget,omitempty"`
+	Recommended        *bool           `json:"recommended,omitempty"`
+	MaxTokens          *int            `json:"maxTokens,omitempty"`
+	MaxOutputTokens    *int            `json:"maxOutputTokens,omitempty"`
+	SupportedMimeTypes map[string]bool `json:"supportedMimeTypes,omitempty"`
+}
+
+// DeprecatedModelInfo 废弃模型转发信息
+type DeprecatedModelInfo struct {
+	NewModelID string `json:"newModelId"`
 }
 
 // FetchAvailableModelsRequest fetchAvailableModels 请求
@@ -524,7 +613,8 @@ type FetchAvailableModelsRequest struct {
 
 // FetchAvailableModelsResponse fetchAvailableModels 响应
 type FetchAvailableModelsResponse struct {
-	Models map[string]ModelInfo `json:"models"`
+	Models             map[string]ModelInfo           `json:"models"`
+	DeprecatedModelIDs map[string]DeprecatedModelInfo `json:"deprecatedModelIds,omitempty"`
 }
 
 // FetchAvailableModels 获取可用模型和配额信息，返回解析后的结构体和原始 JSON
@@ -573,6 +663,13 @@ func (c *Client) FetchAvailableModels(ctx context.Context, accessToken, projectI
 			continue
 		}
 
+		if resp.StatusCode == http.StatusForbidden {
+			return nil, nil, &ForbiddenError{
+				StatusCode: resp.StatusCode,
+				Body:       string(respBodyBytes),
+			}
+		}
+
 		if resp.StatusCode != http.StatusOK {
 			return nil, nil, fmt.Errorf("fetchAvailableModels 失败 (HTTP %d): %s", resp.StatusCode, string(respBodyBytes))
 		}

backend/internal/pkg/antigravity/client_test.go

@@ -190,7 +190,7 @@ func TestTierInfo_UnmarshalJSON_通过JSON嵌套结构(t *testing.T) {
 func TestGetTier_PaidTier优先(t *testing.T) {
 	resp := &LoadCodeAssistResponse{
 		CurrentTier: &TierInfo{ID: "free-tier"},
-		PaidTier:    &TierInfo{ID: "g1-pro-tier"},
+		PaidTier:    &PaidTierInfo{ID: "g1-pro-tier"},
 	}
 	if got := resp.GetTier(); got != "g1-pro-tier" {
 		t.Errorf("应返回 paidTier: got %s", got)
@@ -209,7 +209,7 @@ func TestGetTier_回退到CurrentTier(t *testing.T) {
 func TestGetTier_PaidTier为空ID(t *testing.T) {
 	resp := &LoadCodeAssistResponse{
 		CurrentTier: &TierInfo{ID: "free-tier"},
-		PaidTier:    &TierInfo{ID: ""},
+		PaidTier:    &PaidTierInfo{ID: ""},
 	}
 	// paidTier.ID 为空时应回退到 currentTier
 	if got := resp.GetTier(); got != "free-tier" {
@@ -217,6 +217,32 @@ func TestGetTier_PaidTier为空ID(t *testing.T) {
 	}
 }
 
+func TestGetAvailableCredits(t *testing.T) {
+	resp := &LoadCodeAssistResponse{
+		PaidTier: &PaidTierInfo{
+			ID: "g1-pro-tier",
+			AvailableCredits: []AvailableCredit{
+				{
+					CreditType:                  "GOOGLE_ONE_AI",
+					CreditAmount:                "25",
+					MinimumCreditAmountForUsage: "5",
+				},
+			},
+		},
+	}
+
+	credits := resp.GetAvailableCredits()
+	if len(credits) != 1 {
+		t.Fatalf("AI Credits 数量不匹配: got %d", len(credits))
+	}
+	if credits[0].GetAmount() != 25 {
+		t.Errorf("CreditAmount 解析不正确: got %v", credits[0].GetAmount())
+	}
+	if credits[0].GetMinimumAmount() != 5 {
+		t.Errorf("MinimumCreditAmountForUsage 解析不正确: got %v", credits[0].GetMinimumAmount())
+	}
+}
+
 func TestGetTier_两者都为nil(t *testing.T) {
 	resp := &LoadCodeAssistResponse{}
 	if got := resp.GetTier(); got != "" {

backend/internal/pkg/antigravity/gemini_types.go

@@ -189,6 +189,5 @@ var DefaultStopSequences = []string{
 	"<|user|>",
 	"<|endoftext|>",
 	"<|end_of_turn|>",
-	"[DONE]",
 	"\n\nHuman:",
 }

backend/internal/pkg/antigravity/oauth.go

@@ -49,8 +49,8 @@ const (
 	antigravityDailyBaseURL = "https://daily-cloudcode-pa.sandbox.googleapis.com"
 )
 
-// defaultUserAgentVersion 可通过环境变量 ANTIGRAVITY_USER_AGENT_VERSION 配置，默认 1.20.4
-var defaultUserAgentVersion = "1.20.4"
+// defaultUserAgentVersion 可通过环境变量 ANTIGRAVITY_USER_AGENT_VERSION 配置，默认 1.20.5
+var defaultUserAgentVersion = "1.20.5"
 
 // defaultClientSecret 可通过环境变量 ANTIGRAVITY_OAUTH_CLIENT_SECRET 配置
 var defaultClientSecret = "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"

backend/internal/pkg/antigravity/oauth_test.go

@@ -690,7 +690,7 @@ func TestConstants_值正确(t *testing.T) {
 	if RedirectURI != "http://localhost:8085/callback" {
 		t.Errorf("RedirectURI 不匹配: got %s", RedirectURI)
 	}
-	if GetUserAgent() != "antigravity/1.20.4 windows/amd64" {
+	if GetUserAgent() != "antigravity/1.20.5 windows/amd64" {
 		t.Errorf("UserAgent 不匹配: got %s", GetUserAgent())
 	}
 	if SessionTTL != 30*time.Minute {

backend/internal/pkg/apicompat/anthropic_responses_test.go

@@ -105,6 +105,7 @@ func TestAnthropicToResponses_ToolUse(t *testing.T) {
 	assert.Equal(t, "assistant", items[1].Role)
 	assert.Equal(t, "function_call", items[2].Type)
 	assert.Equal(t, "fc_call_1", items[2].CallID)
+	assert.Empty(t, items[2].ID)
 	assert.Equal(t, "function_call_output", items[3].Type)
 	assert.Equal(t, "fc_call_1", items[3].CallID)
 	assert.Equal(t, "Sunny, 72°F", items[3].Output)

backend/internal/pkg/apicompat/anthropic_to_responses.go

@@ -277,7 +277,6 @@ func anthropicAssistantToResponses(raw json.RawMessage) ([]ResponsesInputItem, e
 			CallID:    fcID,
 			Name:      b.Name,
 			Arguments: args,
-			ID:        fcID,
 		})
 	}
 

backend/internal/pkg/apicompat/chatcompletions_responses_test.go

@@ -99,6 +99,7 @@ func TestChatCompletionsToResponses_ToolCalls(t *testing.T) {
 	// Check function_call item
 	assert.Equal(t, "function_call", items[1].Type)
 	assert.Equal(t, "call_1", items[1].CallID)
+	assert.Empty(t, items[1].ID)
 	assert.Equal(t, "ping", items[1].Name)
 
 	// Check function_call_output item
@@ -252,6 +253,55 @@ func TestChatCompletionsToResponses_AssistantWithTextAndToolCalls(t *testing.T)
 	assert.Equal(t, "user", items[0].Role)
 	assert.Equal(t, "assistant", items[1].Role)
 	assert.Equal(t, "function_call", items[2].Type)
+	assert.Empty(t, items[2].ID)
+}
+
+func TestChatCompletionsToResponses_AssistantArrayContentPreserved(t *testing.T) {
+	req := &ChatCompletionsRequest{
+		Model: "gpt-4o",
+		Messages: []ChatMessage{
+			{Role: "user", Content: json.RawMessage(`"Hi"`)},
+			{Role: "assistant", Content: json.RawMessage(`[{"type":"text","text":"A"},{"type":"text","text":"B"}]`)},
+		},
+	}
+
+	resp, err := ChatCompletionsToResponses(req)
+	require.NoError(t, err)
+
+	var items []ResponsesInputItem
+	require.NoError(t, json.Unmarshal(resp.Input, &items))
+	require.Len(t, items, 2)
+	assert.Equal(t, "assistant", items[1].Role)
+
+	var parts []ResponsesContentPart
+	require.NoError(t, json.Unmarshal(items[1].Content, &parts))
+	require.Len(t, parts, 1)
+	assert.Equal(t, "output_text", parts[0].Type)
+	assert.Equal(t, "AB", parts[0].Text)
+}
+
+func TestChatCompletionsToResponses_AssistantThinkingTagPreserved(t *testing.T) {
+	req := &ChatCompletionsRequest{
+		Model: "gpt-4o",
+		Messages: []ChatMessage{
+			{Role: "user", Content: json.RawMessage(`"Hi"`)},
+			{Role: "assistant", Content: json.RawMessage(`[{"type":"thinking","thinking":"internal plan"},{"type":"text","text":"final answer"}]`)},
+		},
+	}
+
+	resp, err := ChatCompletionsToResponses(req)
+	require.NoError(t, err)
+
+	var items []ResponsesInputItem
+	require.NoError(t, json.Unmarshal(resp.Input, &items))
+	require.Len(t, items, 2)
+
+	var parts []ResponsesContentPart
+	require.NoError(t, json.Unmarshal(items[1].Content, &parts))
+	require.Len(t, parts, 1)
+	assert.Equal(t, "output_text", parts[0].Type)
+	assert.Contains(t, parts[0].Text, "<thinking>internal plan</thinking>")
+	assert.Contains(t, parts[0].Text, "final answer")
 }
 
 // ---------------------------------------------------------------------------
@@ -344,8 +394,8 @@ func TestResponsesToChatCompletions_Reasoning(t *testing.T) {
 
 	var content string
 	require.NoError(t, json.Unmarshal(chat.Choices[0].Message.Content, &content))
-	// Reasoning summary is prepended to text
-	assert.Equal(t, "I thought about it.The answer is 42.", content)
+	assert.Equal(t, "The answer is 42.", content)
+	assert.Equal(t, "I thought about it.", chat.Choices[0].Message.ReasoningContent)
 }
 
 func TestResponsesToChatCompletions_Incomplete(t *testing.T) {
@@ -582,8 +632,35 @@ func TestResponsesEventToChatChunks_ReasoningDelta(t *testing.T) {
 		Delta: "Thinking...",
 	}, state)
 	require.Len(t, chunks, 1)
+	require.NotNil(t, chunks[0].Choices[0].Delta.ReasoningContent)
+	assert.Equal(t, "Thinking...", *chunks[0].Choices[0].Delta.ReasoningContent)
+
+	chunks = ResponsesEventToChatChunks(&ResponsesStreamEvent{
+		Type: "response.reasoning_summary_text.done",
+	}, state)
+	require.Len(t, chunks, 0)
+}
+
+func TestResponsesEventToChatChunks_ReasoningThenTextAutoCloseTag(t *testing.T) {
+	state := NewResponsesEventToChatState()
+	state.Model = "gpt-4o"
+	state.SentRole = true
+
+	chunks := ResponsesEventToChatChunks(&ResponsesStreamEvent{
+		Type:  "response.reasoning_summary_text.delta",
+		Delta: "plan",
+	}, state)
+	require.Len(t, chunks, 1)
+	require.NotNil(t, chunks[0].Choices[0].Delta.ReasoningContent)
+	assert.Equal(t, "plan", *chunks[0].Choices[0].Delta.ReasoningContent)
+
+	chunks = ResponsesEventToChatChunks(&ResponsesStreamEvent{
+		Type:  "response.output_text.delta",
+		Delta: "answer",
+	}, state)
+	require.Len(t, chunks, 1)
 	require.NotNil(t, chunks[0].Choices[0].Delta.Content)
-	assert.Equal(t, "Thinking...", *chunks[0].Choices[0].Delta.Content)
+	assert.Equal(t, "answer", *chunks[0].Choices[0].Delta.Content)
 }
 
 func TestFinalizeResponsesChatStream(t *testing.T) {

backend/internal/pkg/apicompat/chatcompletions_to_responses.go

@@ -3,6 +3,7 @@ package apicompat
 import (
 	"encoding/json"
 	"fmt"
+	"strings"
 )
 
 // ChatCompletionsToResponses converts a Chat Completions request into a
@@ -174,8 +175,11 @@ func chatAssistantToResponses(m ChatMessage) ([]ResponsesInputItem, error) {
 
 	// Emit assistant message with output_text if content is non-empty.
 	if len(m.Content) > 0 {
-		var s string
-		if err := json.Unmarshal(m.Content, &s); err == nil && s != "" {
+		s, err := parseAssistantContent(m.Content)
+		if err != nil {
+			return nil, err
+		}
+		if s != "" {
 			parts := []ResponsesContentPart{{Type: "output_text", Text: s}}
 			partsJSON, err := json.Marshal(parts)
 			if err != nil {
@@ -196,13 +200,82 @@ func chatAssistantToResponses(m ChatMessage) ([]ResponsesInputItem, error) {
 			CallID:    tc.ID,
 			Name:      tc.Function.Name,
 			Arguments: args,
-			ID:        tc.ID,
 		})
 	}
 
 	return items, nil
 }
 
+// parseAssistantContent returns assistant content as plain text.
+//
+// Supported formats:
+// - JSON string
+// - JSON array of typed parts (e.g. [{"type":"text","text":"..."}])
+//
+// For structured thinking/reasoning parts, it preserves semantics by wrapping
+// the text in explicit tags so downstream can still distinguish it from normal text.
+func parseAssistantContent(raw json.RawMessage) (string, error) {
+	if len(raw) == 0 {
+		return "", nil
+	}
+
+	var s string
+	if err := json.Unmarshal(raw, &s); err == nil {
+		return s, nil
+	}
+
+	var parts []map[string]any
+	if err := json.Unmarshal(raw, &parts); err != nil {
+		// Keep compatibility with prior behavior: unsupported assistant content
+		// formats are ignored instead of failing the whole request conversion.
+		return "", nil
+	}
+
+	var b strings.Builder
+	write := func(v string) error {
+		_, err := b.WriteString(v)
+		return err
+	}
+	for _, p := range parts {
+		typ, _ := p["type"].(string)
+		text, _ := p["text"].(string)
+		thinking, _ := p["thinking"].(string)
+
+		switch typ {
+		case "thinking", "reasoning":
+			if thinking != "" {
+				if err := write("<thinking>"); err != nil {
+					return "", err
+				}
+				if err := write(thinking); err != nil {
+					return "", err
+				}
+				if err := write("</thinking>"); err != nil {
+					return "", err
+				}
+			} else if text != "" {
+				if err := write("<thinking>"); err != nil {
+					return "", err
+				}
+				if err := write(text); err != nil {
+					return "", err
+				}
+				if err := write("</thinking>"); err != nil {
+					return "", err
+				}
+			}
+		default:
+			if text != "" {
+				if err := write(text); err != nil {
+					return "", err
+				}
+			}
+		}
+	}
+
+	return b.String(), nil
+}
+
 // chatToolToResponses converts a tool result message (role=tool) into a
 // function_call_output item.
 func chatToolToResponses(m ChatMessage) ([]ResponsesInputItem, error) {

backend/internal/pkg/apicompat/responses_to_chatcompletions.go

@@ -29,6 +29,7 @@ func ResponsesToChatCompletions(resp *ResponsesResponse, model string) *ChatComp
 	}
 
 	var contentText string
+	var reasoningText string
 	var toolCalls []ChatToolCall
 
 	for _, item := range resp.Output {
@@ -51,7 +52,7 @@ func ResponsesToChatCompletions(resp *ResponsesResponse, model string) *ChatComp
 		case "reasoning":
 			for _, s := range item.Summary {
 				if s.Type == "summary_text" && s.Text != "" {
-					contentText += s.Text
+					reasoningText += s.Text
 				}
 			}
 		case "web_search_call":
@@ -67,6 +68,9 @@ func ResponsesToChatCompletions(resp *ResponsesResponse, model string) *ChatComp
 		raw, _ := json.Marshal(contentText)
 		msg.Content = raw
 	}
+	if reasoningText != "" {
+		msg.ReasoningContent = reasoningText
+	}
 
 	finishReason := responsesStatusToChatFinishReason(resp.Status, resp.IncompleteDetails, toolCalls)
 
@@ -153,6 +157,8 @@ func ResponsesEventToChatChunks(evt *ResponsesStreamEvent, state *ResponsesEvent
 		return resToChatHandleFuncArgsDelta(evt, state)
 	case "response.reasoning_summary_text.delta":
 		return resToChatHandleReasoningDelta(evt, state)
+	case "response.reasoning_summary_text.done":
+		return nil
 	case "response.completed", "response.incomplete", "response.failed":
 		return resToChatHandleCompleted(evt, state)
 	default:
@@ -276,8 +282,8 @@ func resToChatHandleReasoningDelta(evt *ResponsesStreamEvent, state *ResponsesEv
 	if evt.Delta == "" {
 		return nil
 	}
-	content := evt.Delta
-	return []ChatCompletionsChunk{makeChatDeltaChunk(state, ChatDelta{Content: &content})}
+	reasoning := evt.Delta
+	return []ChatCompletionsChunk{makeChatDeltaChunk(state, ChatDelta{ReasoningContent: &reasoning})}
 }
 
 func resToChatHandleCompleted(evt *ResponsesStreamEvent, state *ResponsesEventToChatState) []ChatCompletionsChunk {

backend/internal/pkg/apicompat/types.go

@@ -361,11 +361,12 @@ type ChatStreamOptions struct {
 
 // ChatMessage is a single message in the Chat Completions conversation.
 type ChatMessage struct {
-	Role       string          `json:"role"` // "system" | "user" | "assistant" | "tool" | "function"
-	Content    json.RawMessage `json:"content,omitempty"`
-	Name       string          `json:"name,omitempty"`
-	ToolCalls  []ChatToolCall  `json:"tool_calls,omitempty"`
-	ToolCallID string          `json:"tool_call_id,omitempty"`
+	Role             string          `json:"role"` // "system" | "user" | "assistant" | "tool" | "function"
+	Content          json.RawMessage `json:"content,omitempty"`
+	ReasoningContent string          `json:"reasoning_content,omitempty"`
+	Name             string          `json:"name,omitempty"`
+	ToolCalls        []ChatToolCall  `json:"tool_calls,omitempty"`
+	ToolCallID       string          `json:"tool_call_id,omitempty"`
 
 	// Legacy function calling
 	FunctionCall *ChatFunctionCall `json:"function_call,omitempty"`
@@ -466,9 +467,10 @@ type ChatChunkChoice struct {
 
 // ChatDelta carries incremental content in a streaming chunk.
 type ChatDelta struct {
-	Role      string         `json:"role,omitempty"`
-	Content   *string        `json:"content,omitempty"` // pointer: omit when not present, null vs "" matters
-	ToolCalls []ChatToolCall `json:"tool_calls,omitempty"`
+	Role             string         `json:"role,omitempty"`
+	Content          *string        `json:"content,omitempty"` // pointer: omit when not present, null vs "" matters
+	ReasoningContent *string        `json:"reasoning_content,omitempty"`
+	ToolCalls        []ChatToolCall `json:"tool_calls,omitempty"`
 }
 
 // ---------------------------------------------------------------------------

backend/internal/pkg/response/response.go

@@ -47,6 +47,15 @@ func Created(c *gin.Context, data any) {
 	})
 }
 
+// Accepted 返回异步接受响应 (HTTP 202)
+func Accepted(c *gin.Context, data any) {
+	c.JSON(http.StatusAccepted, Response{
+		Code:    0,
+		Message: "accepted",
+		Data:    data,
+	})
+}
+
 // Error 返回错误响应
 func Error(c *gin.Context, statusCode int, message string) {
 	c.JSON(statusCode, Response{

backend/internal/pkg/usagestats/usage_log_types.go

@@ -81,6 +81,22 @@ type ModelStat struct {
 	ActualCost          float64 `json:"actual_cost"` // 实际扣除
 }
 
+// EndpointStat represents usage statistics for a single request endpoint.
+type EndpointStat struct {
+	Endpoint    string  `json:"endpoint"`
+	Requests    int64   `json:"requests"`
+	TotalTokens int64   `json:"total_tokens"`
+	Cost        float64 `json:"cost"`        // 标准计费
+	ActualCost  float64 `json:"actual_cost"` // 实际扣除
+}
+
+// GroupUsageSummary represents today's and cumulative cost for a single group.
+type GroupUsageSummary struct {
+	GroupID   int64   `json:"group_id"`
+	TodayCost float64 `json:"today_cost"`
+	TotalCost float64 `json:"total_cost"`
+}
+
 // GroupStat represents usage statistics for a single group
 type GroupStat struct {
 	GroupID     int64   `json:"group_id"`
@@ -96,12 +112,48 @@ type UserUsageTrendPoint struct {
 	Date       string  `json:"date"`
 	UserID     int64   `json:"user_id"`
 	Email      string  `json:"email"`
+	Username   string  `json:"username"`
 	Requests   int64   `json:"requests"`
 	Tokens     int64   `json:"tokens"`
 	Cost       float64 `json:"cost"`        // 标准计费
 	ActualCost float64 `json:"actual_cost"` // 实际扣除
 }
 
+// UserSpendingRankingItem represents a user spending ranking row.
+type UserSpendingRankingItem struct {
+	UserID     int64   `json:"user_id"`
+	Email      string  `json:"email"`
+	ActualCost float64 `json:"actual_cost"` // 实际扣除
+	Requests   int64   `json:"requests"`
+	Tokens     int64   `json:"tokens"`
+}
+
+// UserSpendingRankingResponse represents ranking rows plus total spend for the time range.
+type UserSpendingRankingResponse struct {
+	Ranking         []UserSpendingRankingItem `json:"ranking"`
+	TotalActualCost float64                   `json:"total_actual_cost"`
+	TotalRequests   int64                     `json:"total_requests"`
+	TotalTokens     int64                     `json:"total_tokens"`
+}
+
+// UserBreakdownItem represents per-user usage breakdown within a dimension (group, model, endpoint).
+type UserBreakdownItem struct {
+	UserID      int64   `json:"user_id"`
+	Email       string  `json:"email"`
+	Requests    int64   `json:"requests"`
+	TotalTokens int64   `json:"total_tokens"`
+	Cost        float64 `json:"cost"`        // 标准计费
+	ActualCost  float64 `json:"actual_cost"` // 实际扣除
+}
+
+// UserBreakdownDimension specifies the dimension to filter for user breakdown.
+type UserBreakdownDimension struct {
+	GroupID      int64  // filter by group_id (>0 to enable)
+	Model        string // filter by model name (non-empty to enable)
+	Endpoint     string // filter by endpoint value (non-empty to enable)
+	EndpointType string // "inbound", "upstream", or "path"
+}
+
 // APIKeyUsageTrendPoint represents API key usage trend data point
 type APIKeyUsageTrendPoint struct {
 	Date     string `json:"date"`
@@ -163,15 +215,18 @@ type UsageLogFilters struct {
 
 // UsageStats represents usage statistics
 type UsageStats struct {
-	TotalRequests     int64    `json:"total_requests"`
-	TotalInputTokens  int64    `json:"total_input_tokens"`
-	TotalOutputTokens int64    `json:"total_output_tokens"`
-	TotalCacheTokens  int64    `json:"total_cache_tokens"`
-	TotalTokens       int64    `json:"total_tokens"`
-	TotalCost         float64  `json:"total_cost"`
-	TotalActualCost   float64  `json:"total_actual_cost"`
-	TotalAccountCost  *float64 `json:"total_account_cost,omitempty"`
-	AverageDurationMs float64  `json:"average_duration_ms"`
+	TotalRequests     int64          `json:"total_requests"`
+	TotalInputTokens  int64          `json:"total_input_tokens"`
+	TotalOutputTokens int64          `json:"total_output_tokens"`
+	TotalCacheTokens  int64          `json:"total_cache_tokens"`
+	TotalTokens       int64          `json:"total_tokens"`
+	TotalCost         float64        `json:"total_cost"`
+	TotalActualCost   float64        `json:"total_actual_cost"`
+	TotalAccountCost  *float64       `json:"total_account_cost,omitempty"`
+	AverageDurationMs float64        `json:"average_duration_ms"`
+	Endpoints         []EndpointStat `json:"endpoints,omitempty"`
+	UpstreamEndpoints []EndpointStat `json:"upstream_endpoints,omitempty"`
+	EndpointPaths     []EndpointStat `json:"endpoint_paths,omitempty"`
 }
 
 // BatchUserUsageStats represents usage stats for a single user
@@ -238,7 +293,9 @@ type AccountUsageSummary struct {
 
 // AccountUsageStatsResponse represents the full usage statistics response for an account
 type AccountUsageStatsResponse struct {
-	History []AccountUsageHistory `json:"history"`
-	Summary AccountUsageSummary   `json:"summary"`
-	Models  []ModelStat           `json:"models"`
+	History           []AccountUsageHistory `json:"history"`
+	Summary           AccountUsageSummary   `json:"summary"`
+	Models            []ModelStat           `json:"models"`
+	Endpoints         []EndpointStat        `json:"endpoints"`
+	UpstreamEndpoints []EndpointStat        `json:"upstream_endpoints"`
 }

backend/internal/repository/account_repo.go

@@ -397,9 +397,9 @@ func (r *accountRepository) Update(ctx context.Context, account *service.Account
 	if err := enqueueSchedulerOutbox(ctx, r.sql, service.SchedulerOutboxEventAccountChanged, &account.ID, nil, buildSchedulerGroupPayload(account.GroupIDs)); err != nil {
 		logger.LegacyPrintf("repository.account", "[SchedulerOutbox] enqueue account update failed: account=%d err=%v", account.ID, err)
 	}
-	if account.Status == service.StatusError || account.Status == service.StatusDisabled || !account.Schedulable {
-		r.syncSchedulerAccountSnapshot(ctx, account.ID)
-	}
+	// 普通账号编辑（如 model_mapping / credentials）也需要立即刷新单账号快照，
+	// 否则网关在 outbox worker 延迟或异常时仍可能读到旧配置。
+	r.syncSchedulerAccountSnapshot(ctx, account.ID)
 	return nil
 }
 
@@ -1727,8 +1727,96 @@ func (r *accountRepository) FindByExtraField(ctx context.Context, key string, va
 // nowUTC is a SQL expression to generate a UTC RFC3339 timestamp string.
 const nowUTC = `to_char(NOW() AT TIME ZONE 'UTC', 'YYYY-MM-DD"T"HH24:MI:SS.US"Z"')`
 
+// dailyExpiredExpr is a SQL expression that evaluates to TRUE when daily quota period has expired.
+// Supports both rolling (24h from start) and fixed (pre-computed reset_at) modes.
+const dailyExpiredExpr = `(
+	CASE WHEN COALESCE(extra->>'quota_daily_reset_mode', 'rolling') = 'fixed'
+	THEN NOW() >= COALESCE((extra->>'quota_daily_reset_at')::timestamptz, '1970-01-01'::timestamptz)
+	ELSE COALESCE((extra->>'quota_daily_start')::timestamptz, '1970-01-01'::timestamptz)
+		+ '24 hours'::interval <= NOW()
+	END
+)`
+
+// weeklyExpiredExpr is a SQL expression that evaluates to TRUE when weekly quota period has expired.
+const weeklyExpiredExpr = `(
+	CASE WHEN COALESCE(extra->>'quota_weekly_reset_mode', 'rolling') = 'fixed'
+	THEN NOW() >= COALESCE((extra->>'quota_weekly_reset_at')::timestamptz, '1970-01-01'::timestamptz)
+	ELSE COALESCE((extra->>'quota_weekly_start')::timestamptz, '1970-01-01'::timestamptz)
+		+ '168 hours'::interval <= NOW()
+	END
+)`
+
+// nextDailyResetAtExpr is a SQL expression to compute the next daily reset_at when a reset occurs.
+// For fixed mode: computes the next future reset time based on NOW(), timezone, and configured hour.
+// This correctly handles long-inactive accounts by jumping directly to the next valid reset point.
+const nextDailyResetAtExpr = `(
+	CASE WHEN COALESCE(extra->>'quota_daily_reset_mode', 'rolling') = 'fixed'
+	THEN to_char((
+		-- Compute today's reset point in the configured timezone, then pick next future one
+		CASE WHEN NOW() >= (
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_daily_reset_hour')::int, 0) || ' hours')::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		-- NOW() is at or past today's reset point → next reset is tomorrow
+		THEN (
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_daily_reset_hour')::int, 0) || ' hours')::interval
+			+ '1 day'::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		-- NOW() is before today's reset point → next reset is today
+		ELSE (
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_daily_reset_hour')::int, 0) || ' hours')::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		END
+	) AT TIME ZONE 'UTC', 'YYYY-MM-DD"T"HH24:MI:SS"Z"')
+	ELSE NULL END
+)`
+
+// nextWeeklyResetAtExpr is a SQL expression to compute the next weekly reset_at when a reset occurs.
+// For fixed mode: computes the next future reset time based on NOW(), timezone, configured day and hour.
+// This correctly handles long-inactive accounts by jumping directly to the next valid reset point.
+const nextWeeklyResetAtExpr = `(
+	CASE WHEN COALESCE(extra->>'quota_weekly_reset_mode', 'rolling') = 'fixed'
+	THEN to_char((
+		-- Compute this week's reset point in the configured timezone
+		-- Step 1: get today's date at reset hour in configured tz
+		-- Step 2: compute days forward to target weekday
+		-- Step 3: if same day but past reset hour, advance 7 days
+		CASE
+		WHEN (
+			-- days_forward = (target_day - current_day + 7) % 7
+			(COALESCE((extra->>'quota_weekly_reset_day')::int, 1)
+			 - EXTRACT(DOW FROM NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))::int
+			 + 7) % 7
+		) = 0 AND NOW() >= (
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_weekly_reset_hour')::int, 0) || ' hours')::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		-- Same weekday and past reset hour → next week
+		THEN (
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_weekly_reset_hour')::int, 0) || ' hours')::interval
+			+ '7 days'::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		ELSE (
+			-- Advance to target weekday this week (or next if days_forward > 0)
+			date_trunc('day', NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))
+			+ (COALESCE((extra->>'quota_weekly_reset_hour')::int, 0) || ' hours')::interval
+			+ ((
+				(COALESCE((extra->>'quota_weekly_reset_day')::int, 1)
+				 - EXTRACT(DOW FROM NOW() AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC'))::int
+				 + 7) % 7
+			) || ' days')::interval
+		) AT TIME ZONE COALESCE(extra->>'quota_reset_timezone', 'UTC')
+		END
+	) AT TIME ZONE 'UTC', 'YYYY-MM-DD"T"HH24:MI:SS"Z"')
+	ELSE NULL END
+)`
+
 // IncrementQuotaUsed 原子递增账号的配额用量（总/日/周三个维度）
 // 日/周额度在周期过期时自动重置为 0 再递增。
+// 支持滚动窗口（rolling）和固定时间（fixed）两种重置模式。
 func (r *accountRepository) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
 	rows, err := r.sql.QueryContext(ctx,
 		`UPDATE accounts SET extra = (
@@ -1739,31 +1827,35 @@ func (r *accountRepository) IncrementQuotaUsed(ctx context.Context, id int64, am
 			|| CASE WHEN COALESCE((extra->>'quota_daily_limit')::numeric, 0) > 0 THEN
 				jsonb_build_object(
 					'quota_daily_used',
-					CASE WHEN COALESCE((extra->>'quota_daily_start')::timestamptz, '1970-01-01'::timestamptz)
-						+ '24 hours'::interval <= NOW()
+					CASE WHEN `+dailyExpiredExpr+`
 					THEN $1
 					ELSE COALESCE((extra->>'quota_daily_used')::numeric, 0) + $1 END,
 					'quota_daily_start',
-					CASE WHEN COALESCE((extra->>'quota_daily_start')::timestamptz, '1970-01-01'::timestamptz)
-						+ '24 hours'::interval <= NOW()
+					CASE WHEN `+dailyExpiredExpr+`
 					THEN `+nowUTC+`
 					ELSE COALESCE(extra->>'quota_daily_start', `+nowUTC+`) END
 				)
+				-- 固定模式重置时更新下次重置时间
+				|| CASE WHEN `+dailyExpiredExpr+` AND `+nextDailyResetAtExpr+` IS NOT NULL
+				   THEN jsonb_build_object('quota_daily_reset_at', `+nextDailyResetAtExpr+`)
+				   ELSE '{}'::jsonb END
 			ELSE '{}'::jsonb END
 			-- 周额度：仅在 quota_weekly_limit > 0 时处理
 			|| CASE WHEN COALESCE((extra->>'quota_weekly_limit')::numeric, 0) > 0 THEN
 				jsonb_build_object(
 					'quota_weekly_used',
-					CASE WHEN COALESCE((extra->>'quota_weekly_start')::timestamptz, '1970-01-01'::timestamptz)
-						+ '168 hours'::interval <= NOW()
+					CASE WHEN `+weeklyExpiredExpr+`
 					THEN $1
 					ELSE COALESCE((extra->>'quota_weekly_used')::numeric, 0) + $1 END,
 					'quota_weekly_start',
-					CASE WHEN COALESCE((extra->>'quota_weekly_start')::timestamptz, '1970-01-01'::timestamptz)
-						+ '168 hours'::interval <= NOW()
+					CASE WHEN `+weeklyExpiredExpr+`
 					THEN `+nowUTC+`
 					ELSE COALESCE(extra->>'quota_weekly_start', `+nowUTC+`) END
 				)
+				-- 固定模式重置时更新下次重置时间
+				|| CASE WHEN `+weeklyExpiredExpr+` AND `+nextWeeklyResetAtExpr+` IS NOT NULL
+				   THEN jsonb_build_object('quota_weekly_reset_at', `+nextWeeklyResetAtExpr+`)
+				   ELSE '{}'::jsonb END
 			ELSE '{}'::jsonb END
 		), updated_at = NOW()
 		WHERE id = $2 AND deleted_at IS NULL
@@ -1796,12 +1888,13 @@ func (r *accountRepository) IncrementQuotaUsed(ctx context.Context, id int64, am
 }
 
 // ResetQuotaUsed 重置账号所有维度的配额用量为 0
+// 保留固定重置模式的配置字段（quota_daily_reset_mode 等），仅清零用量和窗口起始时间
 func (r *accountRepository) ResetQuotaUsed(ctx context.Context, id int64) error {
 	_, err := r.sql.ExecContext(ctx,
 		`UPDATE accounts SET extra = (
 			COALESCE(extra, '{}'::jsonb)
 			|| '{"quota_used": 0, "quota_daily_used": 0, "quota_weekly_used": 0}'::jsonb
-		) - 'quota_daily_start' - 'quota_weekly_start', updated_at = NOW()
+		) - 'quota_daily_start' - 'quota_weekly_start' - 'quota_daily_reset_at' - 'quota_weekly_reset_at', updated_at = NOW()
 		WHERE id = $1 AND deleted_at IS NULL`,
 		id)
 	if err != nil {

backend/internal/repository/account_repo_integration_test.go

@@ -142,6 +142,35 @@ func (s *AccountRepoSuite) TestUpdate_SyncSchedulerSnapshotOnDisabled() {
 	s.Require().Equal(service.StatusDisabled, cacheRecorder.setAccounts[0].Status)
 }
 
+func (s *AccountRepoSuite) TestUpdate_SyncSchedulerSnapshotOnCredentialsChange() {
+	account := mustCreateAccount(s.T(), s.client, &service.Account{
+		Name:        "sync-credentials-update",
+		Status:      service.StatusActive,
+		Schedulable: true,
+		Credentials: map[string]any{
+			"model_mapping": map[string]any{
+				"gpt-5": "gpt-5.1",
+			},
+		},
+	})
+	cacheRecorder := &schedulerCacheRecorder{}
+	s.repo.schedulerCache = cacheRecorder
+
+	account.Credentials = map[string]any{
+		"model_mapping": map[string]any{
+			"gpt-5": "gpt-5.2",
+		},
+	}
+	err := s.repo.Update(s.ctx, account)
+	s.Require().NoError(err, "Update")
+
+	s.Require().Len(cacheRecorder.setAccounts, 1)
+	s.Require().Equal(account.ID, cacheRecorder.setAccounts[0].ID)
+	mapping, ok := cacheRecorder.setAccounts[0].Credentials["model_mapping"].(map[string]any)
+	s.Require().True(ok)
+	s.Require().Equal("gpt-5.2", mapping["gpt-5"])
+}
+
 func (s *AccountRepoSuite) TestDelete() {
 	account := mustCreateAccount(s.T(), s.client, &service.Account{Name: "to-delete"})
 

backend/internal/repository/backup_pg_dumper.go

@@ -0,0 +1,98 @@
+package repository
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"os/exec"
+
+	"github.com/Wei-Shaw/sub2api/internal/config"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+)
+
+// PgDumper implements service.DBDumper using pg_dump/psql
+type PgDumper struct {
+	cfg *config.DatabaseConfig
+}
+
+// NewPgDumper creates a new PgDumper
+func NewPgDumper(cfg *config.Config) service.DBDumper {
+	return &PgDumper{cfg: &cfg.Database}
+}
+
+// Dump executes pg_dump and returns a streaming reader of the output
+func (d *PgDumper) Dump(ctx context.Context) (io.ReadCloser, error) {
+	args := []string{
+		"-h", d.cfg.Host,
+		"-p", fmt.Sprintf("%d", d.cfg.Port),
+		"-U", d.cfg.User,
+		"-d", d.cfg.DBName,
+		"--no-owner",
+		"--no-acl",
+		"--clean",
+		"--if-exists",
+	}
+
+	cmd := exec.CommandContext(ctx, "pg_dump", args...)
+	if d.cfg.Password != "" {
+		cmd.Env = append(cmd.Environ(), "PGPASSWORD="+d.cfg.Password)
+	}
+	if d.cfg.SSLMode != "" {
+		cmd.Env = append(cmd.Environ(), "PGSSLMODE="+d.cfg.SSLMode)
+	}
+
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		return nil, fmt.Errorf("create stdout pipe: %w", err)
+	}
+
+	if err := cmd.Start(); err != nil {
+		return nil, fmt.Errorf("start pg_dump: %w", err)
+	}
+
+	// 返回一个 ReadCloser：读 stdout，关闭时等待进程退出
+	return &cmdReadCloser{ReadCloser: stdout, cmd: cmd}, nil
+}
+
+// Restore executes psql to restore from a streaming reader
+func (d *PgDumper) Restore(ctx context.Context, data io.Reader) error {
+	args := []string{
+		"-h", d.cfg.Host,
+		"-p", fmt.Sprintf("%d", d.cfg.Port),
+		"-U", d.cfg.User,
+		"-d", d.cfg.DBName,
+		"--single-transaction",
+	}
+
+	cmd := exec.CommandContext(ctx, "psql", args...)
+	if d.cfg.Password != "" {
+		cmd.Env = append(cmd.Environ(), "PGPASSWORD="+d.cfg.Password)
+	}
+	if d.cfg.SSLMode != "" {
+		cmd.Env = append(cmd.Environ(), "PGSSLMODE="+d.cfg.SSLMode)
+	}
+
+	cmd.Stdin = data
+
+	output, err := cmd.CombinedOutput()
+	if err != nil {
+		return fmt.Errorf("%v: %s", err, string(output))
+	}
+	return nil
+}
+
+// cmdReadCloser wraps a command stdout pipe and waits for the process on Close
+type cmdReadCloser struct {
+	io.ReadCloser
+	cmd *exec.Cmd
+}
+
+func (c *cmdReadCloser) Close() error {
+	// Close the pipe first
+	_ = c.ReadCloser.Close()
+	// Wait for the process to exit
+	if err := c.cmd.Wait(); err != nil {
+		return fmt.Errorf("pg_dump exited with error: %w", err)
+	}
+	return nil
+}

backend/internal/repository/backup_s3_store.go

@@ -0,0 +1,117 @@
+package repository
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"io"
+	"time"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	v4 "github.com/aws/aws-sdk-go-v2/aws/signer/v4"
+	awsconfig "github.com/aws/aws-sdk-go-v2/config"
+	"github.com/aws/aws-sdk-go-v2/credentials"
+	"github.com/aws/aws-sdk-go-v2/service/s3"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+)
+
+// S3BackupStore implements service.BackupObjectStore using AWS S3 compatible storage
+type S3BackupStore struct {
+	client *s3.Client
+	bucket string
+}
+
+// NewS3BackupStoreFactory returns a BackupObjectStoreFactory that creates S3-backed stores
+func NewS3BackupStoreFactory() service.BackupObjectStoreFactory {
+	return func(ctx context.Context, cfg *service.BackupS3Config) (service.BackupObjectStore, error) {
+		region := cfg.Region
+		if region == "" {
+			region = "auto" // Cloudflare R2 默认 region
+		}
+
+		awsCfg, err := awsconfig.LoadDefaultConfig(ctx,
+			awsconfig.WithRegion(region),
+			awsconfig.WithCredentialsProvider(
+				credentials.NewStaticCredentialsProvider(cfg.AccessKeyID, cfg.SecretAccessKey, ""),
+			),
+		)
+		if err != nil {
+			return nil, fmt.Errorf("load aws config: %w", err)
+		}
+
+		client := s3.NewFromConfig(awsCfg, func(o *s3.Options) {
+			if cfg.Endpoint != "" {
+				o.BaseEndpoint = &cfg.Endpoint
+			}
+			if cfg.ForcePathStyle {
+				o.UsePathStyle = true
+			}
+			o.APIOptions = append(o.APIOptions, v4.SwapComputePayloadSHA256ForUnsignedPayloadMiddleware)
+			o.RequestChecksumCalculation = aws.RequestChecksumCalculationWhenRequired
+		})
+
+		return &S3BackupStore{client: client, bucket: cfg.Bucket}, nil
+	}
+}
+
+func (s *S3BackupStore) Upload(ctx context.Context, key string, body io.Reader, contentType string) (int64, error) {
+	// 读取全部内容以获取大小（S3 PutObject 需要知道内容长度）
+	// 注意：阿里云 OSS 不兼容 s3manager 分片上传的签名方式，因此使用 PutObject
+	data, err := io.ReadAll(body)
+	if err != nil {
+		return 0, fmt.Errorf("read body: %w", err)
+	}
+
+	_, err = s.client.PutObject(ctx, &s3.PutObjectInput{
+		Bucket:      &s.bucket,
+		Key:         &key,
+		Body:        bytes.NewReader(data),
+		ContentType: &contentType,
+	})
+	if err != nil {
+		return 0, fmt.Errorf("S3 PutObject: %w", err)
+	}
+	return int64(len(data)), nil
+}
+
+func (s *S3BackupStore) Download(ctx context.Context, key string) (io.ReadCloser, error) {
+	result, err := s.client.GetObject(ctx, &s3.GetObjectInput{
+		Bucket: &s.bucket,
+		Key:    &key,
+	})
+	if err != nil {
+		return nil, fmt.Errorf("S3 GetObject: %w", err)
+	}
+	return result.Body, nil
+}
+
+func (s *S3BackupStore) Delete(ctx context.Context, key string) error {
+	_, err := s.client.DeleteObject(ctx, &s3.DeleteObjectInput{
+		Bucket: &s.bucket,
+		Key:    &key,
+	})
+	return err
+}
+
+func (s *S3BackupStore) PresignURL(ctx context.Context, key string, expiry time.Duration) (string, error) {
+	presignClient := s3.NewPresignClient(s.client)
+	result, err := presignClient.PresignGetObject(ctx, &s3.GetObjectInput{
+		Bucket: &s.bucket,
+		Key:    &key,
+	}, s3.WithPresignExpires(expiry))
+	if err != nil {
+		return "", fmt.Errorf("presign url: %w", err)
+	}
+	return result.URL, nil
+}
+
+func (s *S3BackupStore) HeadBucket(ctx context.Context) error {
+	_, err := s.client.HeadBucket(ctx, &s3.HeadBucketInput{
+		Bucket: &s.bucket,
+	})
+	if err != nil {
+		return fmt.Errorf("S3 HeadBucket failed: %w", err)
+	}
+	return nil
+}

backend/internal/repository/billing_cache.go

@@ -20,6 +20,11 @@ const (
 	billingCacheTTL           = 5 * time.Minute
 	billingCacheJitter        = 30 * time.Second
 	rateLimitCacheTTL         = 7 * 24 * time.Hour // 7 days matches the longest window
+
+	// Rate limit window durations — must match service.RateLimitWindow* constants.
+	rateLimitWindow5h = 5 * time.Hour
+	rateLimitWindow1d = 24 * time.Hour
+	rateLimitWindow7d = 7 * 24 * time.Hour
 )
 
 // jitteredTTL 返回带随机抖动的 TTL，防止缓存雪崩
@@ -90,17 +95,40 @@ var (
 		return 1
 	`)
 
-	// updateRateLimitUsageScript atomically increments all three rate limit usage counters.
-	// Returns 0 if the key doesn't exist (cache miss), 1 on success.
+	// updateRateLimitUsageScript atomically increments all three rate limit usage counters
+	// with window expiration checking. If a window has expired, its usage is reset to cost
+	// (instead of accumulated) and the window timestamp is updated, matching the DB-side
+	// IncrementRateLimitUsage semantics.
+	//
+	// ARGV: [1]=cost, [2]=ttl_seconds, [3]=now_unix, [4]=window_5h_seconds, [5]=window_1d_seconds, [6]=window_7d_seconds
 	updateRateLimitUsageScript = redis.NewScript(`
 		local exists = redis.call('EXISTS', KEYS[1])
 		if exists == 0 then
 			return 0
 		end
 		local cost = tonumber(ARGV[1])
-		redis.call('HINCRBYFLOAT', KEYS[1], 'usage_5h', cost)
-		redis.call('HINCRBYFLOAT', KEYS[1], 'usage_1d', cost)
-		redis.call('HINCRBYFLOAT', KEYS[1], 'usage_7d', cost)
+		local now = tonumber(ARGV[3])
+		local win5h = tonumber(ARGV[4])
+		local win1d = tonumber(ARGV[5])
+		local win7d = tonumber(ARGV[6])
+
+		-- Helper: check if window is expired and update usage + window accordingly
+		-- Returns nothing, modifies the hash in-place.
+		local function update_window(usage_field, window_field, window_duration)
+			local w = tonumber(redis.call('HGET', KEYS[1], window_field) or 0)
+			if w == 0 or (now - w) >= window_duration then
+				-- Window expired or never started: reset usage to cost, start new window
+				redis.call('HSET', KEYS[1], usage_field, tostring(cost))
+				redis.call('HSET', KEYS[1], window_field, tostring(now))
+			else
+				-- Window still valid: accumulate
+				redis.call('HINCRBYFLOAT', KEYS[1], usage_field, cost)
+			end
+		end
+
+		update_window('usage_5h', 'window_5h', win5h)
+		update_window('usage_1d', 'window_1d', win1d)
+		update_window('usage_7d', 'window_7d', win7d)
 		redis.call('EXPIRE', KEYS[1], ARGV[2])
 		return 1
 	`)
@@ -280,7 +308,15 @@ func (c *billingCache) SetAPIKeyRateLimit(ctx context.Context, keyID int64, data
 
 func (c *billingCache) UpdateAPIKeyRateLimitUsage(ctx context.Context, keyID int64, cost float64) error {
 	key := billingRateLimitKey(keyID)
-	_, err := updateRateLimitUsageScript.Run(ctx, c.rdb, []string{key}, cost, int(rateLimitCacheTTL.Seconds())).Result()
+	now := time.Now().Unix()
+	_, err := updateRateLimitUsageScript.Run(ctx, c.rdb, []string{key},
+		cost,
+		int(rateLimitCacheTTL.Seconds()),
+		now,
+		int(rateLimitWindow5h.Seconds()),
+		int(rateLimitWindow1d.Seconds()),
+		int(rateLimitWindow7d.Seconds()),
+	).Result()
 	if err != nil && !errors.Is(err, redis.Nil) {
 		log.Printf("Warning: update rate limit usage cache failed for api key %d: %v", keyID, err)
 		return err

backend/internal/repository/dashboard_aggregation_repo.go

@@ -17,6 +17,9 @@ type dashboardAggregationRepository struct {
 	sql sqlExecutor
 }
 
+const usageLogsCleanupBatchSize = 10000
+const usageBillingDedupCleanupBatchSize = 10000
+
 // NewDashboardAggregationRepository 创建仪表盘预聚合仓储。
 func NewDashboardAggregationRepository(sqlDB *sql.DB) service.DashboardAggregationRepository {
 	if sqlDB == nil {
@@ -42,6 +45,9 @@ func isPostgresDriver(db *sql.DB) bool {
 }
 
 func (r *dashboardAggregationRepository) AggregateRange(ctx context.Context, start, end time.Time) error {
+	if r == nil || r.sql == nil {
+		return nil
+	}
 	loc := timezone.Location()
 	startLocal := start.In(loc)
 	endLocal := end.In(loc)
@@ -61,6 +67,22 @@ func (r *dashboardAggregationRepository) AggregateRange(ctx context.Context, sta
 		dayEnd = dayEnd.Add(24 * time.Hour)
 	}
 
+	if db, ok := r.sql.(*sql.DB); ok {
+		tx, err := db.BeginTx(ctx, nil)
+		if err != nil {
+			return err
+		}
+		txRepo := newDashboardAggregationRepositoryWithSQL(tx)
+		if err := txRepo.aggregateRangeInTx(ctx, hourStart, hourEnd, dayStart, dayEnd); err != nil {
+			_ = tx.Rollback()
+			return err
+		}
+		return tx.Commit()
+	}
+	return r.aggregateRangeInTx(ctx, hourStart, hourEnd, dayStart, dayEnd)
+}
+
+func (r *dashboardAggregationRepository) aggregateRangeInTx(ctx context.Context, hourStart, hourEnd, dayStart, dayEnd time.Time) error {
 	// 以桶边界聚合，允许覆盖 end 所在桶的剩余区间。
 	if err := r.insertHourlyActiveUsers(ctx, hourStart, hourEnd); err != nil {
 		return err
@@ -195,8 +217,58 @@ func (r *dashboardAggregationRepository) CleanupUsageLogs(ctx context.Context, c
 	if isPartitioned {
 		return r.dropUsageLogsPartitions(ctx, cutoff)
 	}
-	_, err = r.sql.ExecContext(ctx, "DELETE FROM usage_logs WHERE created_at < $1", cutoff.UTC())
-	return err
+	for {
+		res, err := r.sql.ExecContext(ctx, `
+			WITH victims AS (
+				SELECT ctid
+				FROM usage_logs
+				WHERE created_at < $1
+				LIMIT $2
+			)
+			DELETE FROM usage_logs
+			WHERE ctid IN (SELECT ctid FROM victims)
+		`, cutoff.UTC(), usageLogsCleanupBatchSize)
+		if err != nil {
+			return err
+		}
+		affected, err := res.RowsAffected()
+		if err != nil {
+			return err
+		}
+		if affected < usageLogsCleanupBatchSize {
+			return nil
+		}
+	}
+}
+
+func (r *dashboardAggregationRepository) CleanupUsageBillingDedup(ctx context.Context, cutoff time.Time) error {
+	for {
+		res, err := r.sql.ExecContext(ctx, `
+			WITH victims AS (
+				SELECT ctid, request_id, api_key_id, request_fingerprint, created_at
+				FROM usage_billing_dedup
+				WHERE created_at < $1
+				LIMIT $2
+			), archived AS (
+				INSERT INTO usage_billing_dedup_archive (request_id, api_key_id, request_fingerprint, created_at)
+				SELECT request_id, api_key_id, request_fingerprint, created_at
+				FROM victims
+				ON CONFLICT (request_id, api_key_id) DO NOTHING
+			)
+			DELETE FROM usage_billing_dedup
+			WHERE ctid IN (SELECT ctid FROM victims)
+		`, cutoff.UTC(), usageBillingDedupCleanupBatchSize)
+		if err != nil {
+			return err
+		}
+		affected, err := res.RowsAffected()
+		if err != nil {
+			return err
+		}
+		if affected < usageBillingDedupCleanupBatchSize {
+			return nil
+		}
+	}
 }
 
 func (r *dashboardAggregationRepository) EnsureUsageLogsPartitions(ctx context.Context, now time.Time) error {

backend/internal/repository/fixtures_integration_test.go

@@ -262,6 +262,42 @@ func mustCreateApiKey(t *testing.T, client *dbent.Client, k *service.APIKey) *se
 		SetKey(k.Key).
 		SetName(k.Name).
 		SetStatus(k.Status)
+	if k.Quota != 0 {
+		create.SetQuota(k.Quota)
+	}
+	if k.QuotaUsed != 0 {
+		create.SetQuotaUsed(k.QuotaUsed)
+	}
+	if k.RateLimit5h != 0 {
+		create.SetRateLimit5h(k.RateLimit5h)
+	}
+	if k.RateLimit1d != 0 {
+		create.SetRateLimit1d(k.RateLimit1d)
+	}
+	if k.RateLimit7d != 0 {
+		create.SetRateLimit7d(k.RateLimit7d)
+	}
+	if k.Usage5h != 0 {
+		create.SetUsage5h(k.Usage5h)
+	}
+	if k.Usage1d != 0 {
+		create.SetUsage1d(k.Usage1d)
+	}
+	if k.Usage7d != 0 {
+		create.SetUsage7d(k.Usage7d)
+	}
+	if k.Window5hStart != nil {
+		create.SetWindow5hStart(*k.Window5hStart)
+	}
+	if k.Window1dStart != nil {
+		create.SetWindow1dStart(*k.Window1dStart)
+	}
+	if k.Window7dStart != nil {
+		create.SetWindow7dStart(*k.Window7dStart)
+	}
+	if k.ExpiresAt != nil {
+		create.SetExpiresAt(*k.ExpiresAt)
+	}
 	if k.GroupID != nil {
 		create.SetGroupID(*k.GroupID)
 	}

backend/internal/repository/group_repo.go

@@ -88,8 +88,9 @@ func (r *groupRepository) GetByID(ctx context.Context, id int64) (*service.Group
 	if err != nil {
 		return nil, err
 	}
-	count, _ := r.GetAccountCount(ctx, out.ID)
-	out.AccountCount = count
+	total, active, _ := r.GetAccountCount(ctx, out.ID)
+	out.AccountCount = total
+	out.ActiveAccountCount = active
 	return out, nil
 }
 
@@ -256,7 +257,10 @@ func (r *groupRepository) ListWithFilters(ctx context.Context, params pagination
 	counts, err := r.loadAccountCounts(ctx, groupIDs)
 	if err == nil {
 		for i := range outGroups {
-			outGroups[i].AccountCount = counts[outGroups[i].ID]
+			c := counts[outGroups[i].ID]
+			outGroups[i].AccountCount = c.Total
+			outGroups[i].ActiveAccountCount = c.Active
+			outGroups[i].RateLimitedAccountCount = c.RateLimited
 		}
 	}
 
@@ -283,7 +287,10 @@ func (r *groupRepository) ListActive(ctx context.Context) ([]service.Group, erro
 	counts, err := r.loadAccountCounts(ctx, groupIDs)
 	if err == nil {
 		for i := range outGroups {
-			outGroups[i].AccountCount = counts[outGroups[i].ID]
+			c := counts[outGroups[i].ID]
+			outGroups[i].AccountCount = c.Total
+			outGroups[i].ActiveAccountCount = c.Active
+			outGroups[i].RateLimitedAccountCount = c.RateLimited
 		}
 	}
 
@@ -310,7 +317,10 @@ func (r *groupRepository) ListActiveByPlatform(ctx context.Context, platform str
 	counts, err := r.loadAccountCounts(ctx, groupIDs)
 	if err == nil {
 		for i := range outGroups {
-			outGroups[i].AccountCount = counts[outGroups[i].ID]
+			c := counts[outGroups[i].ID]
+			outGroups[i].AccountCount = c.Total
+			outGroups[i].ActiveAccountCount = c.Active
+			outGroups[i].RateLimitedAccountCount = c.RateLimited
 		}
 	}
 
@@ -369,12 +379,20 @@ func (r *groupRepository) ExistsByIDs(ctx context.Context, ids []int64) (map[int
 	return result, nil
 }
 
-func (r *groupRepository) GetAccountCount(ctx context.Context, groupID int64) (int64, error) {
-	var count int64
-	if err := scanSingleRow(ctx, r.sql, "SELECT COUNT(*) FROM account_groups WHERE group_id = $1", []any{groupID}, &count); err != nil {
-		return 0, err
-	}
-	return count, nil
+func (r *groupRepository) GetAccountCount(ctx context.Context, groupID int64) (total int64, active int64, err error) {
+	var rateLimited int64
+	err = scanSingleRow(ctx, r.sql,
+		`SELECT COUNT(*),
+			COUNT(*) FILTER (WHERE a.status = 'active' AND a.schedulable = true),
+			COUNT(*) FILTER (WHERE a.status = 'active' AND (
+				a.rate_limit_reset_at > NOW() OR
+				a.overload_until > NOW() OR
+				a.temp_unschedulable_until > NOW()
+			))
+		FROM account_groups ag JOIN accounts a ON a.id = ag.account_id
+		WHERE ag.group_id = $1`,
+		[]any{groupID}, &total, &active, &rateLimited)
+	return
 }
 
 func (r *groupRepository) DeleteAccountGroupsByGroupID(ctx context.Context, groupID int64) (int64, error) {
@@ -500,15 +518,32 @@ func (r *groupRepository) DeleteCascade(ctx context.Context, id int64) ([]int64,
 	return affectedUserIDs, nil
 }
 
-func (r *groupRepository) loadAccountCounts(ctx context.Context, groupIDs []int64) (counts map[int64]int64, err error) {
-	counts = make(map[int64]int64, len(groupIDs))
+type groupAccountCounts struct {
+	Total       int64
+	Active      int64
+	RateLimited int64
+}
+
+func (r *groupRepository) loadAccountCounts(ctx context.Context, groupIDs []int64) (counts map[int64]groupAccountCounts, err error) {
+	counts = make(map[int64]groupAccountCounts, len(groupIDs))
 	if len(groupIDs) == 0 {
 		return counts, nil
 	}
 
 	rows, err := r.sql.QueryContext(
 		ctx,
-		"SELECT group_id, COUNT(*) FROM account_groups WHERE group_id = ANY($1) GROUP BY group_id",
+		`SELECT ag.group_id,
+			COUNT(*) AS total,
+			COUNT(*) FILTER (WHERE a.status = 'active' AND a.schedulable = true) AS active,
+			COUNT(*) FILTER (WHERE a.status = 'active' AND (
+				a.rate_limit_reset_at > NOW() OR
+				a.overload_until > NOW() OR
+				a.temp_unschedulable_until > NOW()
+			)) AS rate_limited
+		FROM account_groups ag
+		JOIN accounts a ON a.id = ag.account_id
+		WHERE ag.group_id = ANY($1)
+		GROUP BY ag.group_id`,
 		pq.Array(groupIDs),
 	)
 	if err != nil {
@@ -523,11 +558,11 @@ func (r *groupRepository) loadAccountCounts(ctx context.Context, groupIDs []int6
 
 	for rows.Next() {
 		var groupID int64
-		var count int64
-		if err = rows.Scan(&groupID, &count); err != nil {
+		var c groupAccountCounts
+		if err = rows.Scan(&groupID, &c.Total, &c.Active, &c.RateLimited); err != nil {
 			return nil, err
 		}
-		counts[groupID] = count
+		counts[groupID] = c
 	}
 	if err = rows.Err(); err != nil {
 		return nil, err

backend/internal/repository/group_repo_integration_test.go

@@ -603,7 +603,7 @@ func (s *GroupRepoSuite) TestGetAccountCount() {
 	_, err = s.tx.ExecContext(s.ctx, "INSERT INTO account_groups (account_id, group_id, priority, created_at) VALUES ($1, $2, $3, NOW())", a2, group.ID, 2)
 	s.Require().NoError(err)
 
-	count, err := s.repo.GetAccountCount(s.ctx, group.ID)
+	count, _, err := s.repo.GetAccountCount(s.ctx, group.ID)
 	s.Require().NoError(err, "GetAccountCount")
 	s.Require().Equal(int64(2), count)
 }
@@ -619,7 +619,7 @@ func (s *GroupRepoSuite) TestGetAccountCount_Empty() {
 	}
 	s.Require().NoError(s.repo.Create(s.ctx, group))
 
-	count, err := s.repo.GetAccountCount(s.ctx, group.ID)
+	count, _, err := s.repo.GetAccountCount(s.ctx, group.ID)
 	s.Require().NoError(err)
 	s.Require().Zero(count)
 }
@@ -651,7 +651,7 @@ func (s *GroupRepoSuite) TestDeleteAccountGroupsByGroupID() {
 	s.Require().NoError(err, "DeleteAccountGroupsByGroupID")
 	s.Require().Equal(int64(1), affected, "expected 1 affected row")
 
-	count, err := s.repo.GetAccountCount(s.ctx, g.ID)
+	count, _, err := s.repo.GetAccountCount(s.ctx, g.ID)
 	s.Require().NoError(err, "GetAccountCount")
 	s.Require().Equal(int64(0), count, "expected 0 account groups")
 }
@@ -692,7 +692,7 @@ func (s *GroupRepoSuite) TestDeleteAccountGroupsByGroupID_MultipleAccounts() {
 	s.Require().NoError(err)
 	s.Require().Equal(int64(3), affected)
 
-	count, _ := s.repo.GetAccountCount(s.ctx, g.ID)
+	count, _, _ := s.repo.GetAccountCount(s.ctx, g.ID)
 	s.Require().Zero(count)
 }
 

backend/internal/repository/migrations_schema_integration_test.go

@@ -45,6 +45,20 @@ func TestMigrationsRunner_IsIdempotent_AndSchemaIsUpToDate(t *testing.T) {
 	requireColumn(t, tx, "usage_logs", "request_type", "smallint", 0, false)
 	requireColumn(t, tx, "usage_logs", "openai_ws_mode", "boolean", 0, false)
 
+	// usage_billing_dedup: billing idempotency narrow table
+	var usageBillingDedupRegclass sql.NullString
+	require.NoError(t, tx.QueryRowContext(context.Background(), "SELECT to_regclass('public.usage_billing_dedup')").Scan(&usageBillingDedupRegclass))
+	require.True(t, usageBillingDedupRegclass.Valid, "expected usage_billing_dedup table to exist")
+	requireColumn(t, tx, "usage_billing_dedup", "request_fingerprint", "character varying", 64, false)
+	requireIndex(t, tx, "usage_billing_dedup", "idx_usage_billing_dedup_request_api_key")
+	requireIndex(t, tx, "usage_billing_dedup", "idx_usage_billing_dedup_created_at_brin")
+
+	var usageBillingDedupArchiveRegclass sql.NullString
+	require.NoError(t, tx.QueryRowContext(context.Background(), "SELECT to_regclass('public.usage_billing_dedup_archive')").Scan(&usageBillingDedupArchiveRegclass))
+	require.True(t, usageBillingDedupArchiveRegclass.Valid, "expected usage_billing_dedup_archive table to exist")
+	requireColumn(t, tx, "usage_billing_dedup_archive", "request_fingerprint", "character varying", 64, false)
+	requireIndex(t, tx, "usage_billing_dedup_archive", "usage_billing_dedup_archive_pkey")
+
 	// settings table should exist
 	var settingsRegclass sql.NullString
 	require.NoError(t, tx.QueryRowContext(context.Background(), "SELECT to_regclass('public.settings')").Scan(&settingsRegclass))
@@ -75,6 +89,23 @@ func TestMigrationsRunner_IsIdempotent_AndSchemaIsUpToDate(t *testing.T) {
 	requireColumn(t, tx, "user_allowed_groups", "created_at", "timestamp with time zone", 0, false)
 }
 
+func requireIndex(t *testing.T, tx *sql.Tx, table, index string) {
+	t.Helper()
+
+	var exists bool
+	err := tx.QueryRowContext(context.Background(), `
+SELECT EXISTS (
+	SELECT 1
+	FROM pg_indexes
+	WHERE schemaname = 'public'
+	  AND tablename = $1
+	  AND indexname = $2
+)
+`, table, index).Scan(&exists)
+	require.NoError(t, err, "query pg_indexes for %s.%s", table, index)
+	require.True(t, exists, "expected index %s on %s", index, table)
+}
+
 func requireColumn(t *testing.T, tx *sql.Tx, table, column, dataType string, maxLen int, nullable bool) {
 	t.Helper()
 

backend/internal/repository/req_client_pool.go

@@ -73,3 +73,14 @@ func buildReqClientKey(opts reqClientOptions) string {
 		opts.ForceHTTP2,
 	)
 }
+
+// CreatePrivacyReqClient creates an HTTP client for OpenAI privacy settings API
+// This is exported for use by OpenAIPrivacyService
+// Uses Chrome TLS fingerprint impersonation to bypass Cloudflare checks
+func CreatePrivacyReqClient(proxyURL string) (*req.Client, error) {
+	return getSharedReqClient(reqClientOptions{
+		ProxyURL:    proxyURL,
+		Timeout:     30 * time.Second,
+		Impersonate: true, // Enable Chrome TLS fingerprint impersonation
+	})
+}

backend/internal/repository/usage_billing_repo.go

@@ -0,0 +1,308 @@
+package repository
+
+import (
+	"context"
+	"database/sql"
+	"errors"
+	"strings"
+
+	dbent "github.com/Wei-Shaw/sub2api/ent"
+	"github.com/Wei-Shaw/sub2api/internal/pkg/logger"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+)
+
+type usageBillingRepository struct {
+	db *sql.DB
+}
+
+func NewUsageBillingRepository(_ *dbent.Client, sqlDB *sql.DB) service.UsageBillingRepository {
+	return &usageBillingRepository{db: sqlDB}
+}
+
+func (r *usageBillingRepository) Apply(ctx context.Context, cmd *service.UsageBillingCommand) (_ *service.UsageBillingApplyResult, err error) {
+	if cmd == nil {
+		return &service.UsageBillingApplyResult{}, nil
+	}
+	if r == nil || r.db == nil {
+		return nil, errors.New("usage billing repository db is nil")
+	}
+
+	cmd.Normalize()
+	if cmd.RequestID == "" {
+		return nil, service.ErrUsageBillingRequestIDRequired
+	}
+
+	tx, err := r.db.BeginTx(ctx, nil)
+	if err != nil {
+		return nil, err
+	}
+	defer func() {
+		if tx != nil {
+			_ = tx.Rollback()
+		}
+	}()
+
+	applied, err := r.claimUsageBillingKey(ctx, tx, cmd)
+	if err != nil {
+		return nil, err
+	}
+	if !applied {
+		return &service.UsageBillingApplyResult{Applied: false}, nil
+	}
+
+	result := &service.UsageBillingApplyResult{Applied: true}
+	if err := r.applyUsageBillingEffects(ctx, tx, cmd, result); err != nil {
+		return nil, err
+	}
+
+	if err := tx.Commit(); err != nil {
+		return nil, err
+	}
+	tx = nil
+	return result, nil
+}
+
+func (r *usageBillingRepository) claimUsageBillingKey(ctx context.Context, tx *sql.Tx, cmd *service.UsageBillingCommand) (bool, error) {
+	var id int64
+	err := tx.QueryRowContext(ctx, `
+		INSERT INTO usage_billing_dedup (request_id, api_key_id, request_fingerprint)
+		VALUES ($1, $2, $3)
+		ON CONFLICT (request_id, api_key_id) DO NOTHING
+		RETURNING id
+	`, cmd.RequestID, cmd.APIKeyID, cmd.RequestFingerprint).Scan(&id)
+	if errors.Is(err, sql.ErrNoRows) {
+		var existingFingerprint string
+		if err := tx.QueryRowContext(ctx, `
+			SELECT request_fingerprint
+			FROM usage_billing_dedup
+			WHERE request_id = $1 AND api_key_id = $2
+		`, cmd.RequestID, cmd.APIKeyID).Scan(&existingFingerprint); err != nil {
+			return false, err
+		}
+		if strings.TrimSpace(existingFingerprint) != strings.TrimSpace(cmd.RequestFingerprint) {
+			return false, service.ErrUsageBillingRequestConflict
+		}
+		return false, nil
+	}
+	if err != nil {
+		return false, err
+	}
+	var archivedFingerprint string
+	err = tx.QueryRowContext(ctx, `
+		SELECT request_fingerprint
+		FROM usage_billing_dedup_archive
+		WHERE request_id = $1 AND api_key_id = $2
+	`, cmd.RequestID, cmd.APIKeyID).Scan(&archivedFingerprint)
+	if err == nil {
+		if strings.TrimSpace(archivedFingerprint) != strings.TrimSpace(cmd.RequestFingerprint) {
+			return false, service.ErrUsageBillingRequestConflict
+		}
+		return false, nil
+	}
+	if !errors.Is(err, sql.ErrNoRows) {
+		return false, err
+	}
+	return true, nil
+}
+
+func (r *usageBillingRepository) applyUsageBillingEffects(ctx context.Context, tx *sql.Tx, cmd *service.UsageBillingCommand, result *service.UsageBillingApplyResult) error {
+	if cmd.SubscriptionCost > 0 && cmd.SubscriptionID != nil {
+		if err := incrementUsageBillingSubscription(ctx, tx, *cmd.SubscriptionID, cmd.SubscriptionCost); err != nil {
+			return err
+		}
+	}
+
+	if cmd.BalanceCost > 0 {
+		if err := deductUsageBillingBalance(ctx, tx, cmd.UserID, cmd.BalanceCost); err != nil {
+			return err
+		}
+	}
+
+	if cmd.APIKeyQuotaCost > 0 {
+		exhausted, err := incrementUsageBillingAPIKeyQuota(ctx, tx, cmd.APIKeyID, cmd.APIKeyQuotaCost)
+		if err != nil {
+			return err
+		}
+		result.APIKeyQuotaExhausted = exhausted
+	}
+
+	if cmd.APIKeyRateLimitCost > 0 {
+		if err := incrementUsageBillingAPIKeyRateLimit(ctx, tx, cmd.APIKeyID, cmd.APIKeyRateLimitCost); err != nil {
+			return err
+		}
+	}
+
+	if cmd.AccountQuotaCost > 0 && (strings.EqualFold(cmd.AccountType, service.AccountTypeAPIKey) || strings.EqualFold(cmd.AccountType, service.AccountTypeBedrock)) {
+		if err := incrementUsageBillingAccountQuota(ctx, tx, cmd.AccountID, cmd.AccountQuotaCost); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func incrementUsageBillingSubscription(ctx context.Context, tx *sql.Tx, subscriptionID int64, costUSD float64) error {
+	const updateSQL = `
+		UPDATE user_subscriptions us
+		SET
+			daily_usage_usd = us.daily_usage_usd + $1,
+			weekly_usage_usd = us.weekly_usage_usd + $1,
+			monthly_usage_usd = us.monthly_usage_usd + $1,
+			updated_at = NOW()
+		FROM groups g
+		WHERE us.id = $2
+			AND us.deleted_at IS NULL
+			AND us.group_id = g.id
+			AND g.deleted_at IS NULL
+	`
+	res, err := tx.ExecContext(ctx, updateSQL, costUSD, subscriptionID)
+	if err != nil {
+		return err
+	}
+	affected, err := res.RowsAffected()
+	if err != nil {
+		return err
+	}
+	if affected > 0 {
+		return nil
+	}
+	return service.ErrSubscriptionNotFound
+}
+
+func deductUsageBillingBalance(ctx context.Context, tx *sql.Tx, userID int64, amount float64) error {
+	res, err := tx.ExecContext(ctx, `
+		UPDATE users
+		SET balance = balance - $1,
+			updated_at = NOW()
+		WHERE id = $2 AND deleted_at IS NULL
+	`, amount, userID)
+	if err != nil {
+		return err
+	}
+	affected, err := res.RowsAffected()
+	if err != nil {
+		return err
+	}
+	if affected > 0 {
+		return nil
+	}
+	return service.ErrUserNotFound
+}
+
+func incrementUsageBillingAPIKeyQuota(ctx context.Context, tx *sql.Tx, apiKeyID int64, amount float64) (bool, error) {
+	var exhausted bool
+	err := tx.QueryRowContext(ctx, `
+		UPDATE api_keys
+		SET quota_used = quota_used + $1,
+			status = CASE
+				WHEN quota > 0
+					AND status = $3
+					AND quota_used < quota
+					AND quota_used + $1 >= quota
+				THEN $4
+				ELSE status
+			END,
+			updated_at = NOW()
+		WHERE id = $2 AND deleted_at IS NULL
+		RETURNING quota > 0 AND quota_used >= quota AND quota_used - $1 < quota
+	`, amount, apiKeyID, service.StatusAPIKeyActive, service.StatusAPIKeyQuotaExhausted).Scan(&exhausted)
+	if errors.Is(err, sql.ErrNoRows) {
+		return false, service.ErrAPIKeyNotFound
+	}
+	if err != nil {
+		return false, err
+	}
+	return exhausted, nil
+}
+
+func incrementUsageBillingAPIKeyRateLimit(ctx context.Context, tx *sql.Tx, apiKeyID int64, cost float64) error {
+	res, err := tx.ExecContext(ctx, `
+		UPDATE api_keys SET
+			usage_5h = CASE WHEN window_5h_start IS NOT NULL AND window_5h_start + INTERVAL '5 hours' <= NOW() THEN $1 ELSE usage_5h + $1 END,
+			usage_1d = CASE WHEN window_1d_start IS NOT NULL AND window_1d_start + INTERVAL '24 hours' <= NOW() THEN $1 ELSE usage_1d + $1 END,
+			usage_7d = CASE WHEN window_7d_start IS NOT NULL AND window_7d_start + INTERVAL '7 days' <= NOW() THEN $1 ELSE usage_7d + $1 END,
+			window_5h_start = CASE WHEN window_5h_start IS NULL OR window_5h_start + INTERVAL '5 hours' <= NOW() THEN NOW() ELSE window_5h_start END,
+			window_1d_start = CASE WHEN window_1d_start IS NULL OR window_1d_start + INTERVAL '24 hours' <= NOW() THEN date_trunc('day', NOW()) ELSE window_1d_start END,
+			window_7d_start = CASE WHEN window_7d_start IS NULL OR window_7d_start + INTERVAL '7 days' <= NOW() THEN date_trunc('day', NOW()) ELSE window_7d_start END,
+			updated_at = NOW()
+		WHERE id = $2 AND deleted_at IS NULL
+	`, cost, apiKeyID)
+	if err != nil {
+		return err
+	}
+	affected, err := res.RowsAffected()
+	if err != nil {
+		return err
+	}
+	if affected == 0 {
+		return service.ErrAPIKeyNotFound
+	}
+	return nil
+}
+
+func incrementUsageBillingAccountQuota(ctx context.Context, tx *sql.Tx, accountID int64, amount float64) error {
+	rows, err := tx.QueryContext(ctx,
+		`UPDATE accounts SET extra = (
+			COALESCE(extra, '{}'::jsonb)
+			|| jsonb_build_object('quota_used', COALESCE((extra->>'quota_used')::numeric, 0) + $1)
+			|| CASE WHEN COALESCE((extra->>'quota_daily_limit')::numeric, 0) > 0 THEN
+				jsonb_build_object(
+					'quota_daily_used',
+					CASE WHEN COALESCE((extra->>'quota_daily_start')::timestamptz, '1970-01-01'::timestamptz)
+						+ '24 hours'::interval <= NOW()
+					THEN $1
+					ELSE COALESCE((extra->>'quota_daily_used')::numeric, 0) + $1 END,
+					'quota_daily_start',
+					CASE WHEN COALESCE((extra->>'quota_daily_start')::timestamptz, '1970-01-01'::timestamptz)
+						+ '24 hours'::interval <= NOW()
+					THEN `+nowUTC+`
+					ELSE COALESCE(extra->>'quota_daily_start', `+nowUTC+`) END
+				)
+			ELSE '{}'::jsonb END
+			|| CASE WHEN COALESCE((extra->>'quota_weekly_limit')::numeric, 0) > 0 THEN
+				jsonb_build_object(
+					'quota_weekly_used',
+					CASE WHEN COALESCE((extra->>'quota_weekly_start')::timestamptz, '1970-01-01'::timestamptz)
+						+ '168 hours'::interval <= NOW()
+					THEN $1
+					ELSE COALESCE((extra->>'quota_weekly_used')::numeric, 0) + $1 END,
+					'quota_weekly_start',
+					CASE WHEN COALESCE((extra->>'quota_weekly_start')::timestamptz, '1970-01-01'::timestamptz)
+						+ '168 hours'::interval <= NOW()
+					THEN `+nowUTC+`
+					ELSE COALESCE(extra->>'quota_weekly_start', `+nowUTC+`) END
+				)
+			ELSE '{}'::jsonb END
+		), updated_at = NOW()
+		WHERE id = $2 AND deleted_at IS NULL
+		RETURNING
+			COALESCE((extra->>'quota_used')::numeric, 0),
+			COALESCE((extra->>'quota_limit')::numeric, 0)`,
+		amount, accountID)
+	if err != nil {
+		return err
+	}
+	defer func() { _ = rows.Close() }()
+
+	var newUsed, limit float64
+	if rows.Next() {
+		if err := rows.Scan(&newUsed, &limit); err != nil {
+			return err
+		}
+	} else {
+		if err := rows.Err(); err != nil {
+			return err
+		}
+		return service.ErrAccountNotFound
+	}
+	if err := rows.Err(); err != nil {
+		return err
+	}
+	if limit > 0 && newUsed >= limit && (newUsed-amount) < limit {
+		if err := enqueueSchedulerOutbox(ctx, tx, service.SchedulerOutboxEventAccountChanged, &accountID, nil, nil); err != nil {
+			logger.LegacyPrintf("repository.usage_billing", "[SchedulerOutbox] enqueue quota exceeded failed: account=%d err=%v", accountID, err)
+			return err
+		}
+	}
+	return nil
+}

backend/internal/repository/usage_billing_repo_integration_test.go

@@ -0,0 +1,279 @@
+//go:build integration
+
+package repository
+
+import (
+	"context"
+	"fmt"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/require"
+
+	"github.com/Wei-Shaw/sub2api/internal/service"
+)
+
+func TestUsageBillingRepositoryApply_DeduplicatesBalanceBilling(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := NewUsageBillingRepository(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{
+		Email:        fmt.Sprintf("usage-billing-user-%d@example.com", time.Now().UnixNano()),
+		PasswordHash: "hash",
+		Balance:      100,
+	})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{
+		UserID: user.ID,
+		Key:    "sk-usage-billing-" + uuid.NewString(),
+		Name:   "billing",
+		Quota:  1,
+	})
+	account := mustCreateAccount(t, client, &service.Account{
+		Name: "usage-billing-account-" + uuid.NewString(),
+		Type: service.AccountTypeAPIKey,
+	})
+
+	requestID := uuid.NewString()
+	cmd := &service.UsageBillingCommand{
+		RequestID:           requestID,
+		APIKeyID:            apiKey.ID,
+		UserID:              user.ID,
+		AccountID:           account.ID,
+		AccountType:         service.AccountTypeAPIKey,
+		BalanceCost:         1.25,
+		APIKeyQuotaCost:     1.25,
+		APIKeyRateLimitCost: 1.25,
+	}
+
+	result1, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.NotNil(t, result1)
+	require.True(t, result1.Applied)
+	require.True(t, result1.APIKeyQuotaExhausted)
+
+	result2, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.NotNil(t, result2)
+	require.False(t, result2.Applied)
+
+	var balance float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT balance FROM users WHERE id = $1", user.ID).Scan(&balance))
+	require.InDelta(t, 98.75, balance, 0.000001)
+
+	var quotaUsed float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT quota_used FROM api_keys WHERE id = $1", apiKey.ID).Scan(&quotaUsed))
+	require.InDelta(t, 1.25, quotaUsed, 0.000001)
+
+	var usage5h float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT usage_5h FROM api_keys WHERE id = $1", apiKey.ID).Scan(&usage5h))
+	require.InDelta(t, 1.25, usage5h, 0.000001)
+
+	var status string
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT status FROM api_keys WHERE id = $1", apiKey.ID).Scan(&status))
+	require.Equal(t, service.StatusAPIKeyQuotaExhausted, status)
+
+	var dedupCount int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_billing_dedup WHERE request_id = $1 AND api_key_id = $2", requestID, apiKey.ID).Scan(&dedupCount))
+	require.Equal(t, 1, dedupCount)
+}
+
+func TestUsageBillingRepositoryApply_DeduplicatesSubscriptionBilling(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := NewUsageBillingRepository(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{
+		Email:        fmt.Sprintf("usage-billing-sub-user-%d@example.com", time.Now().UnixNano()),
+		PasswordHash: "hash",
+	})
+	group := mustCreateGroup(t, client, &service.Group{
+		Name:             "usage-billing-group-" + uuid.NewString(),
+		Platform:         service.PlatformAnthropic,
+		SubscriptionType: service.SubscriptionTypeSubscription,
+	})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{
+		UserID:  user.ID,
+		GroupID: &group.ID,
+		Key:     "sk-usage-billing-sub-" + uuid.NewString(),
+		Name:    "billing-sub",
+	})
+	subscription := mustCreateSubscription(t, client, &service.UserSubscription{
+		UserID:  user.ID,
+		GroupID: group.ID,
+	})
+
+	requestID := uuid.NewString()
+	cmd := &service.UsageBillingCommand{
+		RequestID:        requestID,
+		APIKeyID:         apiKey.ID,
+		UserID:           user.ID,
+		AccountID:        0,
+		SubscriptionID:   &subscription.ID,
+		SubscriptionCost: 2.5,
+	}
+
+	result1, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.True(t, result1.Applied)
+
+	result2, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.False(t, result2.Applied)
+
+	var dailyUsage float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT daily_usage_usd FROM user_subscriptions WHERE id = $1", subscription.ID).Scan(&dailyUsage))
+	require.InDelta(t, 2.5, dailyUsage, 0.000001)
+}
+
+func TestUsageBillingRepositoryApply_RequestFingerprintConflict(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := NewUsageBillingRepository(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{
+		Email:        fmt.Sprintf("usage-billing-conflict-user-%d@example.com", time.Now().UnixNano()),
+		PasswordHash: "hash",
+		Balance:      100,
+	})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{
+		UserID: user.ID,
+		Key:    "sk-usage-billing-conflict-" + uuid.NewString(),
+		Name:   "billing-conflict",
+	})
+
+	requestID := uuid.NewString()
+	_, err := repo.Apply(ctx, &service.UsageBillingCommand{
+		RequestID:   requestID,
+		APIKeyID:    apiKey.ID,
+		UserID:      user.ID,
+		BalanceCost: 1.25,
+	})
+	require.NoError(t, err)
+
+	_, err = repo.Apply(ctx, &service.UsageBillingCommand{
+		RequestID:   requestID,
+		APIKeyID:    apiKey.ID,
+		UserID:      user.ID,
+		BalanceCost: 2.50,
+	})
+	require.ErrorIs(t, err, service.ErrUsageBillingRequestConflict)
+}
+
+func TestUsageBillingRepositoryApply_UpdatesAccountQuota(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := NewUsageBillingRepository(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{
+		Email:        fmt.Sprintf("usage-billing-account-user-%d@example.com", time.Now().UnixNano()),
+		PasswordHash: "hash",
+	})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{
+		UserID: user.ID,
+		Key:    "sk-usage-billing-account-" + uuid.NewString(),
+		Name:   "billing-account",
+	})
+	account := mustCreateAccount(t, client, &service.Account{
+		Name: "usage-billing-account-quota-" + uuid.NewString(),
+		Type: service.AccountTypeAPIKey,
+		Extra: map[string]any{
+			"quota_limit": 100.0,
+		},
+	})
+
+	_, err := repo.Apply(ctx, &service.UsageBillingCommand{
+		RequestID:        uuid.NewString(),
+		APIKeyID:         apiKey.ID,
+		UserID:           user.ID,
+		AccountID:        account.ID,
+		AccountType:      service.AccountTypeAPIKey,
+		AccountQuotaCost: 3.5,
+	})
+	require.NoError(t, err)
+
+	var quotaUsed float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COALESCE((extra->>'quota_used')::numeric, 0) FROM accounts WHERE id = $1", account.ID).Scan(&quotaUsed))
+	require.InDelta(t, 3.5, quotaUsed, 0.000001)
+}
+
+func TestDashboardAggregationRepositoryCleanupUsageBillingDedup_BatchDeletesOldRows(t *testing.T) {
+	ctx := context.Background()
+	repo := newDashboardAggregationRepositoryWithSQL(integrationDB)
+
+	oldRequestID := "dedup-old-" + uuid.NewString()
+	newRequestID := "dedup-new-" + uuid.NewString()
+	oldCreatedAt := time.Now().UTC().AddDate(0, 0, -400)
+	newCreatedAt := time.Now().UTC().Add(-time.Hour)
+
+	_, err := integrationDB.ExecContext(ctx, `
+		INSERT INTO usage_billing_dedup (request_id, api_key_id, request_fingerprint, created_at)
+		VALUES ($1, 1, $2, $3), ($4, 1, $5, $6)
+	`,
+		oldRequestID, strings.Repeat("a", 64), oldCreatedAt,
+		newRequestID, strings.Repeat("b", 64), newCreatedAt,
+	)
+	require.NoError(t, err)
+
+	require.NoError(t, repo.CleanupUsageBillingDedup(ctx, time.Now().UTC().AddDate(0, 0, -365)))
+
+	var oldCount int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_billing_dedup WHERE request_id = $1", oldRequestID).Scan(&oldCount))
+	require.Equal(t, 0, oldCount)
+
+	var newCount int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_billing_dedup WHERE request_id = $1", newRequestID).Scan(&newCount))
+	require.Equal(t, 1, newCount)
+
+	var archivedCount int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_billing_dedup_archive WHERE request_id = $1", oldRequestID).Scan(&archivedCount))
+	require.Equal(t, 1, archivedCount)
+}
+
+func TestUsageBillingRepositoryApply_DeduplicatesAgainstArchivedKey(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := NewUsageBillingRepository(client, integrationDB)
+	aggRepo := newDashboardAggregationRepositoryWithSQL(integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{
+		Email:        fmt.Sprintf("usage-billing-archive-user-%d@example.com", time.Now().UnixNano()),
+		PasswordHash: "hash",
+		Balance:      100,
+	})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{
+		UserID: user.ID,
+		Key:    "sk-usage-billing-archive-" + uuid.NewString(),
+		Name:   "billing-archive",
+	})
+
+	requestID := uuid.NewString()
+	cmd := &service.UsageBillingCommand{
+		RequestID:   requestID,
+		APIKeyID:    apiKey.ID,
+		UserID:      user.ID,
+		BalanceCost: 1.25,
+	}
+
+	result1, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.True(t, result1.Applied)
+
+	_, err = integrationDB.ExecContext(ctx, `
+		UPDATE usage_billing_dedup
+		SET created_at = $1
+		WHERE request_id = $2 AND api_key_id = $3
+	`, time.Now().UTC().AddDate(0, 0, -400), requestID, apiKey.ID)
+	require.NoError(t, err)
+	require.NoError(t, aggRepo.CleanupUsageBillingDedup(ctx, time.Now().UTC().AddDate(0, 0, -365)))
+
+	result2, err := repo.Apply(ctx, cmd)
+	require.NoError(t, err)
+	require.False(t, result2.Applied)
+
+	var balance float64
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT balance FROM users WHERE id = $1", user.ID).Scan(&balance))
+	require.InDelta(t, 98.75, balance, 0.000001)
+}

backend/internal/repository/usage_log_repo_breakdown_test.go

@@ -0,0 +1,29 @@
+//go:build unit
+
+package repository
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestResolveEndpointColumn(t *testing.T) {
+	tests := []struct {
+		endpointType string
+		want         string
+	}{
+		{"inbound", "ul.inbound_endpoint"},
+		{"upstream", "ul.upstream_endpoint"},
+		{"path", "ul.inbound_endpoint || ' -> ' || ul.upstream_endpoint"},
+		{"", "ul.inbound_endpoint"},           // default
+		{"unknown", "ul.inbound_endpoint"},     // fallback
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.endpointType, func(t *testing.T) {
+			got := resolveEndpointColumn(tc.endpointType)
+			require.Equal(t, tc.want, got)
+		})
+	}
+}

backend/internal/repository/usage_log_repo_integration_test.go

@@ -4,6 +4,8 @@ package repository
 
 import (
 	"context"
+	"fmt"
+	"sync"
 	"testing"
 	"time"
 
@@ -14,6 +16,7 @@ import (
 	"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/usagestats"
 	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/stretchr/testify/require"
 	"github.com/stretchr/testify/suite"
 )
 
@@ -84,6 +87,367 @@ func (s *UsageLogRepoSuite) TestCreate() {
 	s.Require().NotZero(log.ID)
 }
 
+func TestUsageLogRepositoryCreate_BatchPathConcurrent(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-batch-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-batch-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-batch-" + uuid.NewString()})
+
+	const total = 16
+	results := make([]bool, total)
+	errs := make([]error, total)
+	logs := make([]*service.UsageLog, total)
+
+	var wg sync.WaitGroup
+	wg.Add(total)
+	for i := 0; i < total; i++ {
+		i := i
+		logs[i] = &service.UsageLog{
+			UserID:       user.ID,
+			APIKeyID:     apiKey.ID,
+			AccountID:    account.ID,
+			RequestID:    uuid.NewString(),
+			Model:        "claude-3",
+			InputTokens:  10 + i,
+			OutputTokens: 20 + i,
+			TotalCost:    0.5,
+			ActualCost:   0.5,
+			CreatedAt:    time.Now().UTC(),
+		}
+		go func() {
+			defer wg.Done()
+			results[i], errs[i] = repo.Create(ctx, logs[i])
+		}()
+	}
+	wg.Wait()
+
+	for i := 0; i < total; i++ {
+		require.NoError(t, errs[i])
+		require.True(t, results[i])
+		require.NotZero(t, logs[i].ID)
+	}
+
+	var count int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_logs WHERE api_key_id = $1", apiKey.ID).Scan(&count))
+	require.Equal(t, total, count)
+}
+
+func TestUsageLogRepositoryCreate_BatchPathDuplicateRequestID(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-dup-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-dup-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-dup-" + uuid.NewString()})
+	requestID := uuid.NewString()
+
+	log1 := &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    requestID,
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	}
+	log2 := &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    requestID,
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	}
+
+	inserted1, err1 := repo.Create(ctx, log1)
+	inserted2, err2 := repo.Create(ctx, log2)
+	require.NoError(t, err1)
+	require.NoError(t, err2)
+	require.True(t, inserted1)
+	require.False(t, inserted2)
+	require.Equal(t, log1.ID, log2.ID)
+
+	var count int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_logs WHERE request_id = $1 AND api_key_id = $2", requestID, apiKey.ID).Scan(&count))
+	require.Equal(t, 1, count)
+}
+
+func TestUsageLogRepositoryFlushCreateBatch_DeduplicatesSameKeyInMemory(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-batch-memdup-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-batch-memdup-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-batch-memdup-" + uuid.NewString()})
+	requestID := uuid.NewString()
+
+	const total = 8
+	batch := make([]usageLogCreateRequest, 0, total)
+	logs := make([]*service.UsageLog, 0, total)
+
+	for i := 0; i < total; i++ {
+		log := &service.UsageLog{
+			UserID:       user.ID,
+			APIKeyID:     apiKey.ID,
+			AccountID:    account.ID,
+			RequestID:    requestID,
+			Model:        "claude-3",
+			InputTokens:  10 + i,
+			OutputTokens: 20 + i,
+			TotalCost:    0.5,
+			ActualCost:   0.5,
+			CreatedAt:    time.Now().UTC(),
+		}
+		logs = append(logs, log)
+		batch = append(batch, usageLogCreateRequest{
+			log:      log,
+			prepared: prepareUsageLogInsert(log),
+			resultCh: make(chan usageLogCreateResult, 1),
+		})
+	}
+
+	repo.flushCreateBatch(integrationDB, batch)
+
+	insertedCount := 0
+	var firstID int64
+	for idx, req := range batch {
+		res := <-req.resultCh
+		require.NoError(t, res.err)
+		if res.inserted {
+			insertedCount++
+		}
+		require.NotZero(t, logs[idx].ID)
+		if idx == 0 {
+			firstID = logs[idx].ID
+		} else {
+			require.Equal(t, firstID, logs[idx].ID)
+		}
+	}
+
+	require.Equal(t, 1, insertedCount)
+
+	var count int
+	require.NoError(t, integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_logs WHERE request_id = $1 AND api_key_id = $2", requestID, apiKey.ID).Scan(&count))
+	require.Equal(t, 1, count)
+}
+
+func TestUsageLogRepositoryCreateBestEffort_BatchPathDuplicateRequestID(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-best-effort-dup-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-best-effort-dup-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-best-effort-dup-" + uuid.NewString()})
+	requestID := uuid.NewString()
+
+	log1 := &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    requestID,
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	}
+	log2 := &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    requestID,
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	}
+
+	require.NoError(t, repo.CreateBestEffort(ctx, log1))
+	require.NoError(t, repo.CreateBestEffort(ctx, log2))
+
+	require.Eventually(t, func() bool {
+		var count int
+		err := integrationDB.QueryRowContext(ctx, "SELECT COUNT(*) FROM usage_logs WHERE request_id = $1 AND api_key_id = $2", requestID, apiKey.ID).Scan(&count)
+		return err == nil && count == 1
+	}, 3*time.Second, 20*time.Millisecond)
+}
+
+func TestUsageLogRepositoryCreateBestEffort_QueueFullReturnsDropped(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+	repo.bestEffortBatchCh = make(chan usageLogBestEffortRequest, 1)
+	repo.bestEffortBatchCh <- usageLogBestEffortRequest{}
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-best-effort-full-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-best-effort-full-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-best-effort-full-" + uuid.NewString()})
+
+	err := repo.CreateBestEffort(ctx, &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    uuid.NewString(),
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	})
+
+	require.Error(t, err)
+	require.True(t, service.IsUsageLogCreateDropped(err))
+}
+
+func TestUsageLogRepositoryCreate_BatchPathCanceledContextMarksNotPersisted(t *testing.T) {
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-cancel-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-cancel-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-cancel-" + uuid.NewString()})
+
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	inserted, err := repo.Create(ctx, &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    uuid.NewString(),
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	})
+
+	require.False(t, inserted)
+	require.Error(t, err)
+	require.True(t, service.IsUsageLogCreateNotPersisted(err))
+}
+
+func TestUsageLogRepositoryCreate_BatchPathQueueFullMarksNotPersisted(t *testing.T) {
+	ctx := context.Background()
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+	repo.createBatchCh = make(chan usageLogCreateRequest, 1)
+	repo.createBatchCh <- usageLogCreateRequest{}
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-create-full-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-create-full-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-create-full-" + uuid.NewString()})
+
+	inserted, err := repo.Create(ctx, &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    uuid.NewString(),
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	})
+
+	require.False(t, inserted)
+	require.Error(t, err)
+	require.True(t, service.IsUsageLogCreateNotPersisted(err))
+}
+
+func TestUsageLogRepositoryCreate_BatchPathCanceledAfterQueueMarksNotPersisted(t *testing.T) {
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+	repo.createBatchCh = make(chan usageLogCreateRequest, 1)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-cancel-queued-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-cancel-queued-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-cancel-queued-" + uuid.NewString()})
+
+	ctx, cancel := context.WithCancel(context.Background())
+	errCh := make(chan error, 1)
+
+	go func() {
+		_, err := repo.createBatched(ctx, &service.UsageLog{
+			UserID:       user.ID,
+			APIKeyID:     apiKey.ID,
+			AccountID:    account.ID,
+			RequestID:    uuid.NewString(),
+			Model:        "claude-3",
+			InputTokens:  10,
+			OutputTokens: 20,
+			TotalCost:    0.5,
+			ActualCost:   0.5,
+			CreatedAt:    time.Now().UTC(),
+		})
+		errCh <- err
+	}()
+
+	req := <-repo.createBatchCh
+	require.NotNil(t, req.shared)
+	cancel()
+
+	err := <-errCh
+	require.Error(t, err)
+	require.True(t, service.IsUsageLogCreateNotPersisted(err))
+	completeUsageLogCreateRequest(req, usageLogCreateResult{inserted: false, err: service.MarkUsageLogCreateNotPersisted(context.Canceled)})
+}
+
+func TestUsageLogRepositoryFlushCreateBatch_CanceledRequestReturnsNotPersisted(t *testing.T) {
+	client := testEntClient(t)
+	repo := newUsageLogRepositoryWithSQL(client, integrationDB)
+
+	user := mustCreateUser(t, client, &service.User{Email: fmt.Sprintf("usage-flush-cancel-%d@example.com", time.Now().UnixNano())})
+	apiKey := mustCreateApiKey(t, client, &service.APIKey{UserID: user.ID, Key: "sk-usage-flush-cancel-" + uuid.NewString(), Name: "k"})
+	account := mustCreateAccount(t, client, &service.Account{Name: "acc-usage-flush-cancel-" + uuid.NewString()})
+
+	log := &service.UsageLog{
+		UserID:       user.ID,
+		APIKeyID:     apiKey.ID,
+		AccountID:    account.ID,
+		RequestID:    uuid.NewString(),
+		Model:        "claude-3",
+		InputTokens:  10,
+		OutputTokens: 20,
+		TotalCost:    0.5,
+		ActualCost:   0.5,
+		CreatedAt:    time.Now().UTC(),
+	}
+	req := usageLogCreateRequest{
+		log:      log,
+		prepared: prepareUsageLogInsert(log),
+		shared:   &usageLogCreateShared{},
+		resultCh: make(chan usageLogCreateResult, 1),
+	}
+	req.shared.state.Store(usageLogCreateStateCanceled)
+
+	repo.flushCreateBatch(integrationDB, []usageLogCreateRequest{req})
+
+	res := <-req.resultCh
+	require.False(t, res.inserted)
+	require.Error(t, res.err)
+	require.True(t, service.IsUsageLogCreateNotPersisted(res.err))
+}
+
 func (s *UsageLogRepoSuite) TestGetByID() {
 	user := mustCreateUser(s.T(), s.client, &service.User{Email: "getbyid@test.com"})
 	apiKey := mustCreateApiKey(s.T(), s.client, &service.APIKey{UserID: user.ID, Key: "sk-getbyid", Name: "k"})

backend/internal/repository/usage_log_repo_request_type_test.go

@@ -73,6 +73,8 @@ func TestUsageLogRepositoryCreateSyncRequestTypeAndLegacyFields(t *testing.T) {
 			sqlmock.AnyArg(), // media_type
 			sqlmock.AnyArg(), // service_tier
 			sqlmock.AnyArg(), // reasoning_effort
+			sqlmock.AnyArg(), // inbound_endpoint
+			sqlmock.AnyArg(), // upstream_endpoint
 			log.CacheTTLOverridden,
 			createdAt,
 		).
@@ -141,6 +143,8 @@ func TestUsageLogRepositoryCreate_PersistsServiceTier(t *testing.T) {
 			sqlmock.AnyArg(),
 			serviceTier,
 			sqlmock.AnyArg(),
+			sqlmock.AnyArg(),
+			sqlmock.AnyArg(),
 			log.CacheTTLOverridden,
 			createdAt,
 		).
@@ -248,6 +252,37 @@ func TestUsageLogRepositoryGetStatsWithFiltersRequestTypePriority(t *testing.T)
 	require.NoError(t, mock.ExpectationsWereMet())
 }
 
+func TestUsageLogRepositoryGetUserSpendingRanking(t *testing.T) {
+	db, mock := newSQLMock(t)
+	repo := &usageLogRepository{sql: db}
+
+	start := time.Date(2025, 1, 1, 0, 0, 0, 0, time.UTC)
+	end := start.Add(24 * time.Hour)
+
+	rows := sqlmock.NewRows([]string{"user_id", "email", "actual_cost", "requests", "tokens", "total_actual_cost", "total_requests", "total_tokens"}).
+		AddRow(int64(2), "beta@example.com", 12.5, int64(9), int64(900), 40.0, int64(30), int64(2600)).
+		AddRow(int64(1), "alpha@example.com", 12.5, int64(8), int64(800), 40.0, int64(30), int64(2600)).
+		AddRow(int64(3), "gamma@example.com", 4.25, int64(5), int64(300), 40.0, int64(30), int64(2600))
+
+	mock.ExpectQuery("WITH user_spend AS \\(").
+		WithArgs(start, end, 12).
+		WillReturnRows(rows)
+
+	got, err := repo.GetUserSpendingRanking(context.Background(), start, end, 12)
+	require.NoError(t, err)
+	require.Equal(t, &usagestats.UserSpendingRankingResponse{
+		Ranking: []usagestats.UserSpendingRankingItem{
+			{UserID: 2, Email: "beta@example.com", ActualCost: 12.5, Requests: 9, Tokens: 900},
+			{UserID: 1, Email: "alpha@example.com", ActualCost: 12.5, Requests: 8, Tokens: 800},
+			{UserID: 3, Email: "gamma@example.com", ActualCost: 4.25, Requests: 5, Tokens: 300},
+		},
+		TotalActualCost: 40.0,
+		TotalRequests:   30,
+		TotalTokens:     2600,
+	}, got)
+	require.NoError(t, mock.ExpectationsWereMet())
+}
+
 func TestBuildRequestTypeFilterConditionLegacyFallback(t *testing.T) {
 	tests := []struct {
 		name      string
@@ -347,6 +382,8 @@ func TestScanUsageLogRequestTypeAndLegacyFallback(t *testing.T) {
 			sql.NullString{},
 			sql.NullString{Valid: true, String: "priority"},
 			sql.NullString{},
+			sql.NullString{},
+			sql.NullString{},
 			false,
 			now,
 		}})
@@ -386,6 +423,8 @@ func TestScanUsageLogRequestTypeAndLegacyFallback(t *testing.T) {
 			sql.NullString{},
 			sql.NullString{Valid: true, String: "flex"},
 			sql.NullString{},
+			sql.NullString{},
+			sql.NullString{},
 			false,
 			now,
 		}})
@@ -425,6 +464,8 @@ func TestScanUsageLogRequestTypeAndLegacyFallback(t *testing.T) {
 			sql.NullString{},
 			sql.NullString{Valid: true, String: "priority"},
 			sql.NullString{},
+			sql.NullString{},
+			sql.NullString{},
 			false,
 			now,
 		}})

backend/internal/repository/usage_log_repo_unit_test.go

@@ -3,8 +3,11 @@
 package repository
 
 import (
+	"strings"
 	"testing"
+	"time"
 
+	"github.com/Wei-Shaw/sub2api/internal/service"
 	"github.com/stretchr/testify/require"
 )
 
@@ -39,3 +42,26 @@ func TestSafeDateFormat(t *testing.T) {
 		})
 	}
 }
+
+func TestBuildUsageLogBatchInsertQuery_UsesConflictDoNothing(t *testing.T) {
+	log := &service.UsageLog{
+		UserID:       1,
+		APIKeyID:     2,
+		AccountID:    3,
+		RequestID:    "req-batch-no-update",
+		Model:        "gpt-5",
+		InputTokens:  10,
+		OutputTokens: 5,
+		TotalCost:    1.2,
+		ActualCost:   1.2,
+		CreatedAt:    time.Now().UTC(),
+	}
+	prepared := prepareUsageLogInsert(log)
+
+	query, _ := buildUsageLogBatchInsertQuery([]string{usageLogBatchKey(log.RequestID, log.APIKeyID)}, map[string]usageLogInsertPrepared{
+		usageLogBatchKey(log.RequestID, log.APIKeyID): prepared,
+	})
+
+	require.Contains(t, query, "ON CONFLICT (request_id, api_key_id) DO NOTHING")
+	require.NotContains(t, strings.ToUpper(query), "DO UPDATE")
+}

backend/internal/repository/user_group_rate_repo.go

@@ -95,6 +95,35 @@ func (r *userGroupRateRepository) GetByUserIDs(ctx context.Context, userIDs []in
 	return result, nil
 }
 
+// GetByGroupID 获取指定分组下所有用户的专属倍率
+func (r *userGroupRateRepository) GetByGroupID(ctx context.Context, groupID int64) ([]service.UserGroupRateEntry, error) {
+	query := `
+		SELECT ugr.user_id, u.username, u.email, COALESCE(u.notes, ''), u.status, ugr.rate_multiplier
+		FROM user_group_rate_multipliers ugr
+		JOIN users u ON u.id = ugr.user_id
+		WHERE ugr.group_id = $1
+		ORDER BY ugr.user_id
+	`
+	rows, err := r.sql.QueryContext(ctx, query, groupID)
+	if err != nil {
+		return nil, err
+	}
+	defer func() { _ = rows.Close() }()
+
+	var result []service.UserGroupRateEntry
+	for rows.Next() {
+		var entry service.UserGroupRateEntry
+		if err := rows.Scan(&entry.UserID, &entry.UserName, &entry.UserEmail, &entry.UserNotes, &entry.UserStatus, &entry.RateMultiplier); err != nil {
+			return nil, err
+		}
+		result = append(result, entry)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	return result, nil
+}
+
 // GetByUserAndGroup 获取用户在特定分组的专属倍率
 func (r *userGroupRateRepository) GetByUserAndGroup(ctx context.Context, userID, groupID int64) (*float64, error) {
 	query := `SELECT rate_multiplier FROM user_group_rate_multipliers WHERE user_id = $1 AND group_id = $2`
@@ -164,6 +193,31 @@ func (r *userGroupRateRepository) SyncUserGroupRates(ctx context.Context, userID
 	return nil
 }
 
+// SyncGroupRateMultipliers 批量同步分组的用户专属倍率（先删后插）
+func (r *userGroupRateRepository) SyncGroupRateMultipliers(ctx context.Context, groupID int64, entries []service.GroupRateMultiplierInput) error {
+	if _, err := r.sql.ExecContext(ctx, `DELETE FROM user_group_rate_multipliers WHERE group_id = $1`, groupID); err != nil {
+		return err
+	}
+	if len(entries) == 0 {
+		return nil
+	}
+	userIDs := make([]int64, len(entries))
+	rates := make([]float64, len(entries))
+	for i, e := range entries {
+		userIDs[i] = e.UserID
+		rates[i] = e.RateMultiplier
+	}
+	now := time.Now()
+	_, err := r.sql.ExecContext(ctx, `
+		INSERT INTO user_group_rate_multipliers (user_id, group_id, rate_multiplier, created_at, updated_at)
+		SELECT data.user_id, $1::bigint, data.rate_multiplier, $2::timestamptz, $2::timestamptz
+		FROM unnest($3::bigint[], $4::double precision[]) AS data(user_id, rate_multiplier)
+		ON CONFLICT (user_id, group_id)
+		DO UPDATE SET rate_multiplier = EXCLUDED.rate_multiplier, updated_at = EXCLUDED.updated_at
+	`, groupID, now, pq.Array(userIDs), pq.Array(rates))
+	return err
+}
+
 // DeleteByGroupID 删除指定分组的所有用户专属倍率
 func (r *userGroupRateRepository) DeleteByGroupID(ctx context.Context, groupID int64) error {
 	_, err := r.sql.ExecContext(ctx, `DELETE FROM user_group_rate_multipliers WHERE group_id = $1`, groupID)

backend/internal/repository/user_subscription_repo.go

@@ -5,6 +5,7 @@ import (
 	"time"
 
 	dbent "github.com/Wei-Shaw/sub2api/ent"
+	"github.com/Wei-Shaw/sub2api/ent/group"
 	"github.com/Wei-Shaw/sub2api/ent/usersubscription"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/pagination"
 	"github.com/Wei-Shaw/sub2api/internal/service"
@@ -190,7 +191,7 @@ func (r *userSubscriptionRepository) ListByGroupID(ctx context.Context, groupID
 	return userSubscriptionEntitiesToService(subs), paginationResultFromTotal(int64(total), params), nil
 }
 
-func (r *userSubscriptionRepository) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
+func (r *userSubscriptionRepository) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, platform, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	client := clientFromContext(ctx, r.client)
 	q := client.UserSubscription.Query()
 	if userID != nil {
@@ -199,6 +200,9 @@ func (r *userSubscriptionRepository) List(ctx context.Context, params pagination
 	if groupID != nil {
 		q = q.Where(usersubscription.GroupIDEQ(*groupID))
 	}
+	if platform != "" {
+		q = q.Where(usersubscription.HasGroupWith(group.PlatformEQ(platform)))
+	}
 
 	// Status filtering with real-time expiration check
 	now := time.Now()

backend/internal/repository/user_subscription_repo_integration_test.go

@@ -271,7 +271,7 @@ func (s *UserSubscriptionRepoSuite) TestList_NoFilters() {
 	group := s.mustCreateGroup("g-list")
 	s.mustCreateSubscription(user.ID, group.ID, nil)
 
-	subs, page, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, nil, "", "", "")
+	subs, page, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, nil, "", "", "", "")
 	s.Require().NoError(err, "List")
 	s.Require().Len(subs, 1)
 	s.Require().Equal(int64(1), page.Total)
@@ -285,7 +285,7 @@ func (s *UserSubscriptionRepoSuite) TestList_FilterByUserID() {
 	s.mustCreateSubscription(user1.ID, group.ID, nil)
 	s.mustCreateSubscription(user2.ID, group.ID, nil)
 
-	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, &user1.ID, nil, "", "", "")
+	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, &user1.ID, nil, "", "", "", "")
 	s.Require().NoError(err)
 	s.Require().Len(subs, 1)
 	s.Require().Equal(user1.ID, subs[0].UserID)
@@ -299,7 +299,7 @@ func (s *UserSubscriptionRepoSuite) TestList_FilterByGroupID() {
 	s.mustCreateSubscription(user.ID, g1.ID, nil)
 	s.mustCreateSubscription(user.ID, g2.ID, nil)
 
-	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, &g1.ID, "", "", "")
+	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, &g1.ID, "", "", "", "")
 	s.Require().NoError(err)
 	s.Require().Len(subs, 1)
 	s.Require().Equal(g1.ID, subs[0].GroupID)
@@ -320,7 +320,7 @@ func (s *UserSubscriptionRepoSuite) TestList_FilterByStatus() {
 		c.SetExpiresAt(time.Now().Add(-24 * time.Hour))
 	})
 
-	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, nil, service.SubscriptionStatusExpired, "", "")
+	subs, _, err := s.repo.List(s.ctx, pagination.PaginationParams{Page: 1, PageSize: 10}, nil, nil, service.SubscriptionStatusExpired, "", "", "")
 	s.Require().NoError(err)
 	s.Require().Len(subs, 1)
 	s.Require().Equal(service.SubscriptionStatusExpired, subs[0].Status)

backend/internal/repository/wire.go

@@ -62,6 +62,7 @@ var ProviderSet = wire.NewSet(
 	NewAnnouncementRepository,
 	NewAnnouncementReadRepository,
 	NewUsageLogRepository,
+	NewUsageBillingRepository,
 	NewIdempotencyRepository,
 	NewUsageCleanupRepository,
 	NewDashboardAggregationRepository,
@@ -99,6 +100,10 @@ var ProviderSet = wire.NewSet(
 	// Encryptors
 	NewAESEncryptor,
 
+	// Backup infrastructure
+	NewPgDumper,
+	NewS3BackupStoreFactory,
+
 	// HTTP service ports (DI Strategy A: return interface directly)
 	NewTurnstileVerifier,
 	ProvidePricingRemoteClient,

backend/internal/server/api_contract_test.go

@@ -493,6 +493,7 @@ func TestAPIContracts(t *testing.T) {
 					"registration_email_suffix_whitelist": [],
 					"promo_code_enabled": true,
 					"password_reset_enabled": false,
+					"frontend_url": "",
 					"totp_enabled": false,
 					"totp_encryption_key_configured": false,
 					"smtp_host": "smtp.example.com",
@@ -537,6 +538,7 @@ func TestAPIContracts(t *testing.T) {
 					"purchase_subscription_url": "",
 					"min_claude_code_version": "",
 					"allow_ungrouped_key_scheduling": false,
+					"backend_mode_enabled": false,
 					"custom_menu_items": []
 				}
 			}`,
@@ -645,7 +647,7 @@ func newContractDeps(t *testing.T) *contractDeps {
 	settingRepo := newStubSettingRepo()
 	settingService := service.NewSettingService(settingRepo, cfg)
 
-	adminService := service.NewAdminService(userRepo, groupRepo, &accountRepo, nil, proxyRepo, apiKeyRepo, redeemRepo, nil, nil, nil, nil, nil, nil, nil, nil, nil)
+	adminService := service.NewAdminService(userRepo, groupRepo, &accountRepo, nil, proxyRepo, apiKeyRepo, redeemRepo, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil)
 	authHandler := handler.NewAuthHandler(cfg, nil, userService, settingService, nil, redeemService, nil)
 	apiKeyHandler := handler.NewAPIKeyHandler(apiKeyService)
 	usageHandler := handler.NewUsageHandler(usageService, apiKeyService)
@@ -922,8 +924,8 @@ func (stubGroupRepo) ExistsByName(ctx context.Context, name string) (bool, error
 	return false, errors.New("not implemented")
 }
 
-func (stubGroupRepo) GetAccountCount(ctx context.Context, groupID int64) (int64, error) {
-	return 0, errors.New("not implemented")
+func (stubGroupRepo) GetAccountCount(ctx context.Context, groupID int64) (int64, int64, error) {
+	return 0, 0, errors.New("not implemented")
 }
 
 func (stubGroupRepo) DeleteAccountGroupsByGroupID(ctx context.Context, groupID int64) (int64, error) {
@@ -1287,7 +1289,7 @@ func (r *stubUserSubscriptionRepo) ListActiveByUserID(ctx context.Context, userI
 func (stubUserSubscriptionRepo) ListByGroupID(ctx context.Context, groupID int64, params pagination.PaginationParams) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	return nil, nil, errors.New("not implemented")
 }
-func (stubUserSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
+func (stubUserSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, platform, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	return nil, nil, errors.New("not implemented")
 }
 func (stubUserSubscriptionRepo) ExistsByUserIDAndGroupID(ctx context.Context, userID, groupID int64) (bool, error) {
@@ -1623,10 +1625,22 @@ func (r *stubUsageLogRepo) GetModelStatsWithFilters(ctx context.Context, startTi
 	return nil, errors.New("not implemented")
 }
 
+func (r *stubUsageLogRepo) GetEndpointStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, model string, requestType *int16, stream *bool, billingType *int8) ([]usagestats.EndpointStat, error) {
+	return nil, errors.New("not implemented")
+}
+
+func (r *stubUsageLogRepo) GetUpstreamEndpointStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, model string, requestType *int16, stream *bool, billingType *int8) ([]usagestats.EndpointStat, error) {
+	return nil, errors.New("not implemented")
+}
+
 func (r *stubUsageLogRepo) GetGroupStatsWithFilters(ctx context.Context, startTime, endTime time.Time, userID, apiKeyID, accountID, groupID int64, requestType *int16, stream *bool, billingType *int8) ([]usagestats.GroupStat, error) {
 	return nil, errors.New("not implemented")
 }
 
+func (r *stubUsageLogRepo) GetUserBreakdownStats(ctx context.Context, startTime, endTime time.Time, dim usagestats.UserBreakdownDimension, limit int) ([]usagestats.UserBreakdownItem, error) {
+	return nil, errors.New("not implemented")
+}
+
 func (r *stubUsageLogRepo) GetAPIKeyUsageTrend(ctx context.Context, startTime, endTime time.Time, granularity string, limit int) ([]usagestats.APIKeyUsageTrendPoint, error) {
 	return nil, errors.New("not implemented")
 }
@@ -1635,6 +1649,10 @@ func (r *stubUsageLogRepo) GetUserUsageTrend(ctx context.Context, startTime, end
 	return nil, errors.New("not implemented")
 }
 
+func (r *stubUsageLogRepo) GetUserSpendingRanking(ctx context.Context, startTime, endTime time.Time, limit int) (*usagestats.UserSpendingRankingResponse, error) {
+	return nil, errors.New("not implemented")
+}
+
 func (r *stubUsageLogRepo) GetUserStatsAggregated(ctx context.Context, userID int64, startTime, endTime time.Time) (*usagestats.UsageStats, error) {
 	logs := r.userLogs[userID]
 	if len(logs) == 0 {
@@ -1768,6 +1786,9 @@ func (r *stubUsageLogRepo) GetAccountUsageStats(ctx context.Context, accountID i
 func (r *stubUsageLogRepo) GetStatsWithFilters(ctx context.Context, filters usagestats.UsageLogFilters) (*usagestats.UsageStats, error) {
 	return nil, errors.New("not implemented")
 }
+func (r *stubUsageLogRepo) GetAllGroupUsageSummary(ctx context.Context, todayStart time.Time) ([]usagestats.GroupUsageSummary, error) {
+	return nil, errors.New("not implemented")
+}
 
 type stubSettingRepo struct {
 	all map[string]string

backend/internal/server/middleware/api_key_auth_google_test.go

@@ -135,7 +135,7 @@ func (f fakeGoogleSubscriptionRepo) ListActiveByUserID(ctx context.Context, user
 func (f fakeGoogleSubscriptionRepo) ListByGroupID(ctx context.Context, groupID int64, params pagination.PaginationParams) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	return nil, nil, errors.New("not implemented")
 }
-func (f fakeGoogleSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
+func (f fakeGoogleSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, platform, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	return nil, nil, errors.New("not implemented")
 }
 func (f fakeGoogleSubscriptionRepo) ExistsByUserIDAndGroupID(ctx context.Context, userID, groupID int64) (bool, error) {

backend/internal/server/middleware/api_key_auth_test.go

@@ -646,7 +646,7 @@ func (r *stubUserSubscriptionRepo) ListByGroupID(ctx context.Context, groupID in
 	return nil, nil, errors.New("not implemented")
 }
 
-func (r *stubUserSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
+func (r *stubUserSubscriptionRepo) List(ctx context.Context, params pagination.PaginationParams, userID, groupID *int64, status, platform, sortBy, sortOrder string) ([]service.UserSubscription, *pagination.PaginationResult, error) {
 	return nil, nil, errors.New("not implemented")
 }
 

backend/internal/server/middleware/backend_mode_guard.go

@@ -0,0 +1,51 @@
+package middleware
+
+import (
+	"strings"
+
+	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+// BackendModeUserGuard blocks non-admin users from accessing user routes when backend mode is enabled.
+// Must be placed AFTER JWT auth middleware so that the user role is available in context.
+func BackendModeUserGuard(settingService *service.SettingService) gin.HandlerFunc {
+	return func(c *gin.Context) {
+		if settingService == nil || !settingService.IsBackendModeEnabled(c.Request.Context()) {
+			c.Next()
+			return
+		}
+		role, _ := GetUserRoleFromContext(c)
+		if role == "admin" {
+			c.Next()
+			return
+		}
+		response.Forbidden(c, "Backend mode is active. User self-service is disabled.")
+		c.Abort()
+	}
+}
+
+// BackendModeAuthGuard selectively blocks auth endpoints when backend mode is enabled.
+// Allows: login, login/2fa, logout, refresh (admin needs these).
+// Blocks: register, forgot-password, reset-password, OAuth, etc.
+func BackendModeAuthGuard(settingService *service.SettingService) gin.HandlerFunc {
+	return func(c *gin.Context) {
+		if settingService == nil || !settingService.IsBackendModeEnabled(c.Request.Context()) {
+			c.Next()
+			return
+		}
+		path := c.Request.URL.Path
+		// Allow login, 2FA, logout, refresh, public settings
+		allowedSuffixes := []string{"/auth/login", "/auth/login/2fa", "/auth/logout", "/auth/refresh"}
+		for _, suffix := range allowedSuffixes {
+			if strings.HasSuffix(path, suffix) {
+				c.Next()
+				return
+			}
+		}
+		response.Forbidden(c, "Backend mode is active. Registration and self-service auth flows are disabled.")
+		c.Abort()
+	}
+}

backend/internal/server/middleware/backend_mode_guard_test.go

@@ -0,0 +1,239 @@
+//go:build unit
+
+package middleware
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/config"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+	"github.com/gin-gonic/gin"
+	"github.com/stretchr/testify/require"
+)
+
+type bmSettingRepo struct {
+	values map[string]string
+}
+
+func (r *bmSettingRepo) Get(_ context.Context, _ string) (*service.Setting, error) {
+	panic("unexpected Get call")
+}
+
+func (r *bmSettingRepo) GetValue(_ context.Context, key string) (string, error) {
+	v, ok := r.values[key]
+	if !ok {
+		return "", service.ErrSettingNotFound
+	}
+	return v, nil
+}
+
+func (r *bmSettingRepo) Set(_ context.Context, _, _ string) error {
+	panic("unexpected Set call")
+}
+
+func (r *bmSettingRepo) GetMultiple(_ context.Context, _ []string) (map[string]string, error) {
+	panic("unexpected GetMultiple call")
+}
+
+func (r *bmSettingRepo) SetMultiple(_ context.Context, settings map[string]string) error {
+	if r.values == nil {
+		r.values = make(map[string]string, len(settings))
+	}
+	for key, value := range settings {
+		r.values[key] = value
+	}
+	return nil
+}
+
+func (r *bmSettingRepo) GetAll(_ context.Context) (map[string]string, error) {
+	panic("unexpected GetAll call")
+}
+
+func (r *bmSettingRepo) Delete(_ context.Context, _ string) error {
+	panic("unexpected Delete call")
+}
+
+func newBackendModeSettingService(t *testing.T, enabled string) *service.SettingService {
+	t.Helper()
+
+	repo := &bmSettingRepo{
+		values: map[string]string{
+			service.SettingKeyBackendModeEnabled: enabled,
+		},
+	}
+	svc := service.NewSettingService(repo, &config.Config{})
+	require.NoError(t, svc.UpdateSettings(context.Background(), &service.SystemSettings{
+		BackendModeEnabled: enabled == "true",
+	}))
+
+	return svc
+}
+
+func stringPtr(v string) *string {
+	return &v
+}
+
+func TestBackendModeUserGuard(t *testing.T) {
+	tests := []struct {
+		name       string
+		nilService bool
+		enabled    string
+		role       *string
+		wantStatus int
+	}{
+		{
+			name:       "disabled_allows_all",
+			enabled:    "false",
+			role:       stringPtr("user"),
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "nil_service_allows_all",
+			nilService: true,
+			role:       stringPtr("user"),
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_admin_allowed",
+			enabled:    "true",
+			role:       stringPtr("admin"),
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_user_blocked",
+			enabled:    "true",
+			role:       stringPtr("user"),
+			wantStatus: http.StatusForbidden,
+		},
+		{
+			name:       "enabled_no_role_blocked",
+			enabled:    "true",
+			wantStatus: http.StatusForbidden,
+		},
+		{
+			name:       "enabled_empty_role_blocked",
+			enabled:    "true",
+			role:       stringPtr(""),
+			wantStatus: http.StatusForbidden,
+		},
+	}
+
+	for _, tc := range tests {
+		tc := tc
+		t.Run(tc.name, func(t *testing.T) {
+			gin.SetMode(gin.TestMode)
+
+			r := gin.New()
+			if tc.role != nil {
+				role := *tc.role
+				r.Use(func(c *gin.Context) {
+					c.Set(string(ContextKeyUserRole), role)
+					c.Next()
+				})
+			}
+
+			var svc *service.SettingService
+			if !tc.nilService {
+				svc = newBackendModeSettingService(t, tc.enabled)
+			}
+
+			r.Use(BackendModeUserGuard(svc))
+			r.GET("/test", func(c *gin.Context) {
+				c.JSON(http.StatusOK, gin.H{"ok": true})
+			})
+
+			w := httptest.NewRecorder()
+			req := httptest.NewRequest(http.MethodGet, "/test", nil)
+			r.ServeHTTP(w, req)
+
+			require.Equal(t, tc.wantStatus, w.Code)
+		})
+	}
+}
+
+func TestBackendModeAuthGuard(t *testing.T) {
+	tests := []struct {
+		name       string
+		nilService bool
+		enabled    string
+		path       string
+		wantStatus int
+	}{
+		{
+			name:       "disabled_allows_all",
+			enabled:    "false",
+			path:       "/api/v1/auth/register",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "nil_service_allows_all",
+			nilService: true,
+			path:       "/api/v1/auth/register",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_allows_login",
+			enabled:    "true",
+			path:       "/api/v1/auth/login",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_allows_login_2fa",
+			enabled:    "true",
+			path:       "/api/v1/auth/login/2fa",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_allows_logout",
+			enabled:    "true",
+			path:       "/api/v1/auth/logout",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_allows_refresh",
+			enabled:    "true",
+			path:       "/api/v1/auth/refresh",
+			wantStatus: http.StatusOK,
+		},
+		{
+			name:       "enabled_blocks_register",
+			enabled:    "true",
+			path:       "/api/v1/auth/register",
+			wantStatus: http.StatusForbidden,
+		},
+		{
+			name:       "enabled_blocks_forgot_password",
+			enabled:    "true",
+			path:       "/api/v1/auth/forgot-password",
+			wantStatus: http.StatusForbidden,
+		},
+	}
+
+	for _, tc := range tests {
+		tc := tc
+		t.Run(tc.name, func(t *testing.T) {
+			gin.SetMode(gin.TestMode)
+
+			r := gin.New()
+
+			var svc *service.SettingService
+			if !tc.nilService {
+				svc = newBackendModeSettingService(t, tc.enabled)
+			}
+
+			r.Use(BackendModeAuthGuard(svc))
+			r.Any("/*path", func(c *gin.Context) {
+				c.JSON(http.StatusOK, gin.H{"ok": true})
+			})
+
+			w := httptest.NewRecorder()
+			req := httptest.NewRequest(http.MethodGet, tc.path, nil)
+			r.ServeHTTP(w, req)
+
+			require.Equal(t, tc.wantStatus, w.Code)
+		})
+	}
+}

backend/internal/server/router.go

@@ -107,9 +107,9 @@ func registerRoutes(
 	v1 := r.Group("/api/v1")
 
 	// 注册各模块路由
-	routes.RegisterAuthRoutes(v1, h, jwtAuth, redisClient)
-	routes.RegisterUserRoutes(v1, h, jwtAuth)
-	routes.RegisterSoraClientRoutes(v1, h, jwtAuth)
+	routes.RegisterAuthRoutes(v1, h, jwtAuth, redisClient, settingService)
+	routes.RegisterUserRoutes(v1, h, jwtAuth, settingService)
+	routes.RegisterSoraClientRoutes(v1, h, jwtAuth, settingService)
 	routes.RegisterAdminRoutes(v1, h, adminAuth)
 	routes.RegisterGatewayRoutes(r, h, apiKeyAuth, apiKeyService, subscriptionService, opsService, settingService, cfg)
 }

backend/internal/server/routes/admin.go

@@ -58,6 +58,9 @@ func RegisterAdminRoutes(
 		// 数据管理
 		registerDataManagementRoutes(admin, h)
 
+		// 数据库备份恢复
+		registerBackupRoutes(admin, h)
+
 		// 运维监控（Ops）
 		registerOpsRoutes(admin, h)
 
@@ -192,8 +195,10 @@ func registerDashboardRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
 		dashboard.GET("/groups", h.Admin.Dashboard.GetGroupStats)
 		dashboard.GET("/api-keys-trend", h.Admin.Dashboard.GetAPIKeyUsageTrend)
 		dashboard.GET("/users-trend", h.Admin.Dashboard.GetUserUsageTrend)
+		dashboard.GET("/users-ranking", h.Admin.Dashboard.GetUserSpendingRanking)
 		dashboard.POST("/users-usage", h.Admin.Dashboard.GetBatchUsersUsage)
 		dashboard.POST("/api-keys-usage", h.Admin.Dashboard.GetBatchAPIKeysUsage)
+		dashboard.GET("/user-breakdown", h.Admin.Dashboard.GetUserBreakdown)
 		dashboard.POST("/aggregation/backfill", h.Admin.Dashboard.BackfillAggregation)
 	}
 }
@@ -222,12 +227,17 @@ func registerGroupRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
 	{
 		groups.GET("", h.Admin.Group.List)
 		groups.GET("/all", h.Admin.Group.GetAll)
+		groups.GET("/usage-summary", h.Admin.Group.GetUsageSummary)
+		groups.GET("/capacity-summary", h.Admin.Group.GetCapacitySummary)
 		groups.PUT("/sort-order", h.Admin.Group.UpdateSortOrder)
 		groups.GET("/:id", h.Admin.Group.GetByID)
 		groups.POST("", h.Admin.Group.Create)
 		groups.PUT("/:id", h.Admin.Group.Update)
 		groups.DELETE("/:id", h.Admin.Group.Delete)
 		groups.GET("/:id/stats", h.Admin.Group.GetStats)
+		groups.GET("/:id/rate-multipliers", h.Admin.Group.GetGroupRateMultipliers)
+		groups.PUT("/:id/rate-multipliers", h.Admin.Group.BatchSetGroupRateMultipliers)
+		groups.DELETE("/:id/rate-multipliers", h.Admin.Group.ClearGroupRateMultipliers)
 		groups.GET("/:id/api-keys", h.Admin.Group.GetGroupAPIKeys)
 	}
 }
@@ -436,6 +446,30 @@ func registerDataManagementRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
 	}
 }
 
+func registerBackupRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
+	backup := admin.Group("/backups")
+	{
+		// S3 存储配置
+		backup.GET("/s3-config", h.Admin.Backup.GetS3Config)
+		backup.PUT("/s3-config", h.Admin.Backup.UpdateS3Config)
+		backup.POST("/s3-config/test", h.Admin.Backup.TestS3Connection)
+
+		// 定时备份配置
+		backup.GET("/schedule", h.Admin.Backup.GetSchedule)
+		backup.PUT("/schedule", h.Admin.Backup.UpdateSchedule)
+
+		// 备份操作
+		backup.POST("", h.Admin.Backup.CreateBackup)
+		backup.GET("", h.Admin.Backup.ListBackups)
+		backup.GET("/:id", h.Admin.Backup.GetBackup)
+		backup.DELETE("/:id", h.Admin.Backup.DeleteBackup)
+		backup.GET("/:id/download-url", h.Admin.Backup.GetDownloadURL)
+
+		// 恢复操作
+		backup.POST("/:id/restore", h.Admin.Backup.RestoreBackup)
+	}
+}
+
 func registerSystemRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
 	system := admin.Group("/system")
 	{

backend/internal/server/routes/auth.go

@@ -6,6 +6,7 @@ import (
 	"github.com/Wei-Shaw/sub2api/internal/handler"
 	"github.com/Wei-Shaw/sub2api/internal/middleware"
 	servermiddleware "github.com/Wei-Shaw/sub2api/internal/server/middleware"
+	"github.com/Wei-Shaw/sub2api/internal/service"
 
 	"github.com/gin-gonic/gin"
 	"github.com/redis/go-redis/v9"
@@ -17,12 +18,14 @@ func RegisterAuthRoutes(
 	h *handler.Handlers,
 	jwtAuth servermiddleware.JWTAuthMiddleware,
 	redisClient *redis.Client,
+	settingService *service.SettingService,
 ) {
 	// 创建速率限制器
 	rateLimiter := middleware.NewRateLimiter(redisClient)
 
 	// 公开接口
 	auth := v1.Group("/auth")
+	auth.Use(servermiddleware.BackendModeAuthGuard(settingService))
 	{
 		// 注册/登录/2FA/验证码发送均属于高风险入口，增加服务端兜底限流（Redis 故障时 fail-close）
 		auth.POST("/register", rateLimiter.LimitWithOptions("auth-register", 5, time.Minute, middleware.RateLimitOptions{
@@ -78,6 +81,7 @@ func RegisterAuthRoutes(
 	// 需要认证的当前用户信息
 	authenticated := v1.Group("")
 	authenticated.Use(gin.HandlerFunc(jwtAuth))
+	authenticated.Use(servermiddleware.BackendModeUserGuard(settingService))
 	{
 		authenticated.GET("/auth/me", h.Auth.GetCurrentUser)
 		// 撤销所有会话（需要认证）

backend/internal/server/routes/auth_rate_limit_test.go

@@ -29,6 +29,7 @@ func newAuthRoutesTestRouter(redisClient *redis.Client) *gin.Engine {
 			c.Next()
 		}),
 		redisClient,
+		nil,
 	)
 
 	return router

backend/internal/server/routes/gateway.go

@@ -30,6 +30,7 @@ func RegisterGatewayRoutes(
 	soraBodyLimit := middleware.RequestBodyLimit(soraMaxBodySize)
 	clientRequestID := middleware.ClientRequestID()
 	opsErrorLogger := handler.OpsErrorLoggerMiddleware(opsService)
+	endpointNorm := handler.InboundEndpointMiddleware()
 
 	// 未分组 Key 拦截中间件（按协议格式区分错误响应）
 	requireGroupAnthropic := middleware.RequireGroupAssignment(settingService, middleware.AnthropicErrorWriter)
@@ -40,6 +41,7 @@ func RegisterGatewayRoutes(
 	gateway.Use(bodyLimit)
 	gateway.Use(clientRequestID)
 	gateway.Use(opsErrorLogger)
+	gateway.Use(endpointNorm)
 	gateway.Use(gin.HandlerFunc(apiKeyAuth))
 	gateway.Use(requireGroupAnthropic)
 	{
@@ -80,6 +82,7 @@ func RegisterGatewayRoutes(
 	gemini.Use(bodyLimit)
 	gemini.Use(clientRequestID)
 	gemini.Use(opsErrorLogger)
+	gemini.Use(endpointNorm)
 	gemini.Use(middleware.APIKeyAuthWithSubscriptionGoogle(apiKeyService, subscriptionService, cfg))
 	gemini.Use(requireGroupGoogle)
 	{
@@ -90,11 +93,11 @@ func RegisterGatewayRoutes(
 	}
 
 	// OpenAI Responses API（不带v1前缀的别名）
-	r.POST("/responses", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
-	r.POST("/responses/*subpath", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
-	r.GET("/responses", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.ResponsesWebSocket)
+	r.POST("/responses", bodyLimit, clientRequestID, opsErrorLogger, endpointNorm, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
+	r.POST("/responses/*subpath", bodyLimit, clientRequestID, opsErrorLogger, endpointNorm, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
+	r.GET("/responses", bodyLimit, clientRequestID, opsErrorLogger, endpointNorm, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.ResponsesWebSocket)
 	// OpenAI Chat Completions API（不带v1前缀的别名）
-	r.POST("/chat/completions", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.ChatCompletions)
+	r.POST("/chat/completions", bodyLimit, clientRequestID, opsErrorLogger, endpointNorm, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.ChatCompletions)
 
 	// Antigravity 模型列表
 	r.GET("/antigravity/models", gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.Gateway.AntigravityModels)
@@ -104,6 +107,7 @@ func RegisterGatewayRoutes(
 	antigravityV1.Use(bodyLimit)
 	antigravityV1.Use(clientRequestID)
 	antigravityV1.Use(opsErrorLogger)
+	antigravityV1.Use(endpointNorm)
 	antigravityV1.Use(middleware.ForcePlatform(service.PlatformAntigravity))
 	antigravityV1.Use(gin.HandlerFunc(apiKeyAuth))
 	antigravityV1.Use(requireGroupAnthropic)
@@ -118,6 +122,7 @@ func RegisterGatewayRoutes(
 	antigravityV1Beta.Use(bodyLimit)
 	antigravityV1Beta.Use(clientRequestID)
 	antigravityV1Beta.Use(opsErrorLogger)
+	antigravityV1Beta.Use(endpointNorm)
 	antigravityV1Beta.Use(middleware.ForcePlatform(service.PlatformAntigravity))
 	antigravityV1Beta.Use(middleware.APIKeyAuthWithSubscriptionGoogle(apiKeyService, subscriptionService, cfg))
 	antigravityV1Beta.Use(requireGroupGoogle)
@@ -132,6 +137,7 @@ func RegisterGatewayRoutes(
 	soraV1.Use(soraBodyLimit)
 	soraV1.Use(clientRequestID)
 	soraV1.Use(opsErrorLogger)
+	soraV1.Use(endpointNorm)
 	soraV1.Use(middleware.ForcePlatform(service.PlatformSora))
 	soraV1.Use(gin.HandlerFunc(apiKeyAuth))
 	soraV1.Use(requireGroupAnthropic)

backend/internal/server/routes/sora_client.go

@@ -3,6 +3,7 @@ package routes
 import (
 	"github.com/Wei-Shaw/sub2api/internal/handler"
 	"github.com/Wei-Shaw/sub2api/internal/server/middleware"
+	"github.com/Wei-Shaw/sub2api/internal/service"
 
 	"github.com/gin-gonic/gin"
 )
@@ -12,6 +13,7 @@ func RegisterSoraClientRoutes(
 	v1 *gin.RouterGroup,
 	h *handler.Handlers,
 	jwtAuth middleware.JWTAuthMiddleware,
+	settingService *service.SettingService,
 ) {
 	if h.SoraClient == nil {
 		return
@@ -19,6 +21,7 @@ func RegisterSoraClientRoutes(
 
 	authenticated := v1.Group("/sora")
 	authenticated.Use(gin.HandlerFunc(jwtAuth))
+	authenticated.Use(middleware.BackendModeUserGuard(settingService))
 	{
 		authenticated.POST("/generate", h.SoraClient.Generate)
 		authenticated.GET("/generations", h.SoraClient.ListGenerations)