feat: add quota limit for API key accounts

- Add configurable spending limit (quota_limit) for apikey-type accounts
- Atomic quota accumulation via PostgreSQL JSONB operations on TotalCost
- Scheduler filters out over-quota accounts with outbox-triggered snapshot refresh
- Display quota usage ($used / $limit) in account capacity column
- Add "Reset Quota" action in account menu to reset usage to zero
- Editing account settings preserves quota_used (no accidental reset)
- Covers all 3 billing paths: Anthropic, Gemini, OpenAI RecordUsage

chore: bump version to 0.1.90.4

backend/internal/handler/admin/account_handler.go

@@ -1328,6 +1328,29 @@ func (h *AccountHandler) ClearRateLimit(c *gin.Context) {
 	response.Success(c, h.buildAccountResponseWithRuntime(c.Request.Context(), account))
 }
 
+// ResetQuota handles resetting account quota usage
+// POST /api/v1/admin/accounts/:id/reset-quota
+func (h *AccountHandler) ResetQuota(c *gin.Context) {
+	accountID, err := strconv.ParseInt(c.Param("id"), 10, 64)
+	if err != nil {
+		response.BadRequest(c, "Invalid account ID")
+		return
+	}
+
+	if err := h.adminService.ResetAccountQuota(c.Request.Context(), accountID); err != nil {
+		response.InternalError(c, "Failed to reset account quota: "+err.Error())
+		return
+	}
+
+	account, err := h.adminService.GetAccount(c.Request.Context(), accountID)
+	if err != nil {
+		response.ErrorFrom(c, err)
+		return
+	}
+
+	response.Success(c, h.buildAccountResponseWithRuntime(c.Request.Context(), account))
+}
+
 // GetTempUnschedulable handles getting temporary unschedulable status
 // GET /api/v1/admin/accounts/:id/temp-unschedulable
 func (h *AccountHandler) GetTempUnschedulable(c *gin.Context) {

backend/internal/handler/admin/admin_service_stub_test.go

@@ -425,5 +425,9 @@ func (s *stubAdminService) AdminUpdateAPIKeyGroupID(ctx context.Context, keyID i
 	return nil, service.ErrAPIKeyNotFound
 }
 
+func (s *stubAdminService) ResetAccountQuota(ctx context.Context, id int64) error {
+	return nil
+}
+
 // Ensure stub implements interface.
 var _ service.AdminService = (*stubAdminService)(nil)

backend/internal/handler/dto/mappers.go

@@ -248,6 +248,17 @@ 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 {
+		if limit := a.GetQuotaLimit(); limit > 0 {
+			out.QuotaLimit = &limit
+		}
+		used := a.GetQuotaUsed()
+		if out.QuotaLimit != nil {
+			out.QuotaUsed = &used
+		}
+	}
+
 	return out
 }
 

backend/internal/handler/dto/types.go

@@ -185,6 +185,10 @@ type Account struct {
 	CacheTTLOverrideEnabled *bool   `json:"cache_ttl_override_enabled,omitempty"`
 	CacheTTLOverrideTarget  *string `json:"cache_ttl_override_target,omitempty"`
 
+	// API Key 账号配额限制
+	QuotaLimit *float64 `json:"quota_limit,omitempty"`
+	QuotaUsed  *float64 `json:"quota_used,omitempty"`
+
 	Proxy         *Proxy         `json:"proxy,omitempty"`
 	AccountGroups []AccountGroup `json:"account_groups,omitempty"`
 

backend/internal/handler/sora_client_handler_test.go

@@ -2132,6 +2132,14 @@ func (r *stubAccountRepoForHandler) BulkUpdate(context.Context, []int64, service
 	return 0, nil
 }
 
+func (r *stubAccountRepoForHandler) IncrementQuotaUsed(context.Context, int64, float64) error {
+	return nil
+}
+
+func (r *stubAccountRepoForHandler) ResetQuotaUsed(context.Context, int64) error {
+	return nil
+}
+
 // ==================== Stub: SoraClient (用于 SoraGatewayService) ====================
 
 var _ service.SoraClient = (*stubSoraClientForHandler)(nil)

backend/internal/handler/sora_gateway_handler_test.go

@@ -216,6 +216,14 @@ func (r *stubAccountRepo) BulkUpdate(ctx context.Context, ids []int64, updates s
 	return 0, nil
 }
 
+func (r *stubAccountRepo) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
+	return nil
+}
+
+func (r *stubAccountRepo) ResetQuotaUsed(ctx context.Context, id int64) error {
+	return nil
+}
+
 func (r *stubAccountRepo) listSchedulable() []service.Account {
 	var result []service.Account
 	for _, acc := range r.accounts {

backend/internal/repository/account_repo.go

@@ -1657,3 +1657,60 @@ func (r *accountRepository) FindByExtraField(ctx context.Context, key string, va
 
 	return r.accountsToService(ctx, accounts)
 }
+
+// IncrementQuotaUsed 原子递增账号的 extra.quota_used 字段
+func (r *accountRepository) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
+	rows, err := r.sql.QueryContext(ctx,
+		`UPDATE accounts SET extra = jsonb_set(
+			COALESCE(extra, '{}'::jsonb),
+			'{quota_used}',
+			to_jsonb(COALESCE((extra->>'quota_used')::numeric, 0) + $1)
+		), 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, id)
+	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
+		}
+	}
+	if err := rows.Err(); err != nil {
+		return err
+	}
+
+	// 配额刚超限时触发调度快照刷新，使账号及时从调度候选中移除
+	if limit > 0 && newUsed >= limit && (newUsed-amount) < limit {
+		if err := enqueueSchedulerOutbox(ctx, r.sql, service.SchedulerOutboxEventAccountChanged, &id, nil, nil); err != nil {
+			logger.LegacyPrintf("repository.account", "[SchedulerOutbox] enqueue quota exceeded failed: account=%d err=%v", id, err)
+		}
+	}
+	return nil
+}
+
+// ResetQuotaUsed 重置账号的 extra.quota_used 为 0
+func (r *accountRepository) ResetQuotaUsed(ctx context.Context, id int64) error {
+	_, err := r.sql.ExecContext(ctx,
+		`UPDATE accounts SET extra = jsonb_set(
+			COALESCE(extra, '{}'::jsonb),
+			'{quota_used}',
+			'0'::jsonb
+		), updated_at = NOW()
+		WHERE id = $1 AND deleted_at IS NULL`,
+		id)
+	if err != nil {
+		return err
+	}
+	// 重置配额后触发调度快照刷新，使账号重新参与调度
+	if err := enqueueSchedulerOutbox(ctx, r.sql, service.SchedulerOutboxEventAccountChanged, &id, nil, nil); err != nil {
+		logger.LegacyPrintf("repository.account", "[SchedulerOutbox] enqueue quota reset failed: account=%d err=%v", id, err)
+	}
+	return nil
+}

backend/internal/server/routes/admin.go

@@ -252,6 +252,7 @@ func registerAccountRoutes(admin *gin.RouterGroup, h *handler.Handlers) {
 		accounts.GET("/:id/today-stats", h.Admin.Account.GetTodayStats)
 		accounts.POST("/today-stats/batch", h.Admin.Account.GetBatchTodayStats)
 		accounts.POST("/:id/clear-rate-limit", h.Admin.Account.ClearRateLimit)
+		accounts.POST("/:id/reset-quota", h.Admin.Account.ResetQuota)
 		accounts.GET("/:id/temp-unschedulable", h.Admin.Account.GetTempUnschedulable)
 		accounts.DELETE("/:id/temp-unschedulable", h.Admin.Account.ClearTempUnschedulable)
 		accounts.POST("/:id/schedulable", h.Admin.Account.SetSchedulable)

backend/internal/service/account.go

@@ -1117,6 +1117,38 @@ func (a *Account) GetCacheTTLOverrideTarget() string {
 	return "5m"
 }
 
+// GetQuotaLimit 获取 API Key 账号的配额限制（美元）
+// 返回 0 表示未启用
+func (a *Account) GetQuotaLimit() float64 {
+	if a.Extra == nil {
+		return 0
+	}
+	if v, ok := a.Extra["quota_limit"]; ok {
+		return parseExtraFloat64(v)
+	}
+	return 0
+}
+
+// GetQuotaUsed 获取 API Key 账号的已用配额（美元）
+func (a *Account) GetQuotaUsed() float64 {
+	if a.Extra == nil {
+		return 0
+	}
+	if v, ok := a.Extra["quota_used"]; ok {
+		return parseExtraFloat64(v)
+	}
+	return 0
+}
+
+// IsQuotaExceeded 检查 API Key 账号配额是否已超限
+func (a *Account) IsQuotaExceeded() bool {
+	limit := a.GetQuotaLimit()
+	if limit <= 0 {
+		return false
+	}
+	return a.GetQuotaUsed() >= limit
+}
+
 // GetWindowCostLimit 获取 5h 窗口费用阈值（美元）
 // 返回 0 表示未启用
 func (a *Account) GetWindowCostLimit() float64 {

backend/internal/service/account_service.go

@@ -68,6 +68,10 @@ type AccountRepository interface {
 	UpdateSessionWindow(ctx context.Context, id int64, start, end *time.Time, status string) error
 	UpdateExtra(ctx context.Context, id int64, updates map[string]any) error
 	BulkUpdate(ctx context.Context, ids []int64, updates AccountBulkUpdate) (int64, error)
+	// IncrementQuotaUsed 原子递增 API Key 账号的配额用量
+	IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error
+	// ResetQuotaUsed 重置 API Key 账号的配额用量为 0
+	ResetQuotaUsed(ctx context.Context, id int64) error
 }
 
 // AccountBulkUpdate describes the fields that can be updated in a bulk operation.

backend/internal/service/account_service_delete_test.go

@@ -199,6 +199,14 @@ func (s *accountRepoStub) BulkUpdate(ctx context.Context, ids []int64, updates A
 	panic("unexpected BulkUpdate call")
 }
 
+func (s *accountRepoStub) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
+	return nil
+}
+
+func (s *accountRepoStub) ResetQuotaUsed(ctx context.Context, id int64) error {
+	return nil
+}
+
 // TestAccountService_Delete_NotFound 测试删除不存在的账号时返回正确的错误。
 // 预期行为：
 //   - ExistsByID 返回 false（账号不存在）

backend/internal/service/admin_service.go

@@ -84,6 +84,7 @@ type AdminService interface {
 	DeleteRedeemCode(ctx context.Context, id int64) error
 	BatchDeleteRedeemCodes(ctx context.Context, ids []int64) (int64, error)
 	ExpireRedeemCode(ctx context.Context, id int64) (*RedeemCode, error)
+	ResetAccountQuota(ctx context.Context, id int64) error
 }
 
 // CreateUserInput represents input for creating a new user via admin operations.
@@ -1458,6 +1459,10 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U
 		account.Credentials = input.Credentials
 	}
 	if len(input.Extra) > 0 {
+		// 保留 quota_used，防止编辑账号时意外重置配额用量
+		if oldQuotaUsed, ok := account.Extra["quota_used"]; ok {
+			input.Extra["quota_used"] = oldQuotaUsed
+		}
 		account.Extra = input.Extra
 	}
 	if input.ProxyID != nil {
@@ -2439,3 +2444,7 @@ func (e *MixedChannelError) Error() string {
 	return fmt.Sprintf("mixed_channel_warning: Group '%s' contains both %s and %s accounts. Using mixed channels in the same context may cause thinking block signature validation issues, which will fallback to non-thinking mode for historical messages.",
 		e.GroupName, e.CurrentPlatform, e.OtherPlatform)
 }
+
+func (s *adminServiceImpl) ResetAccountQuota(ctx context.Context, id int64) error {
+	return s.accountRepo.ResetQuotaUsed(ctx, id)
+}

backend/internal/service/gateway_multiplatform_test.go

@@ -187,6 +187,14 @@ func (m *mockAccountRepoForPlatform) BulkUpdate(ctx context.Context, ids []int64
 	return 0, nil
 }
 
+func (m *mockAccountRepoForPlatform) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
+	return nil
+}
+
+func (m *mockAccountRepoForPlatform) ResetQuotaUsed(ctx context.Context, id int64) error {
+	return nil
+}
+
 // Verify interface implementation
 var _ AccountRepository = (*mockAccountRepoForPlatform)(nil)
 

backend/internal/service/gateway_service.go

@@ -1228,6 +1228,10 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro
 				modelScopeSkippedIDs = append(modelScopeSkippedIDs, account.ID)
 				continue
 			}
+			// 配额检查
+			if !s.isAccountSchedulableForQuota(account) {
+				continue
+			}
 			// 窗口费用检查（非粘性会话路径）
 			if !s.isAccountSchedulableForWindowCost(ctx, account, false) {
 				filteredWindowCost++
@@ -1260,6 +1264,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro
 							s.isAccountAllowedForPlatform(stickyAccount, platform, useMixed) &&
 							(requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, stickyAccount, requestedModel)) &&
 							s.isAccountSchedulableForModelSelection(ctx, stickyAccount, requestedModel) &&
+							s.isAccountSchedulableForQuota(stickyAccount) &&
 							s.isAccountSchedulableForWindowCost(ctx, stickyAccount, true) &&
 
 							s.isAccountSchedulableForRPM(ctx, stickyAccount, true) { // 粘性会话窗口费用+RPM 检查
@@ -1416,6 +1421,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro
 					s.isAccountAllowedForPlatform(account, platform, useMixed) &&
 					(requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) &&
 					s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) &&
+					s.isAccountSchedulableForQuota(account) &&
 					s.isAccountSchedulableForWindowCost(ctx, account, true) &&
 
 					s.isAccountSchedulableForRPM(ctx, account, true) { // 粘性会话窗口费用+RPM 检查
@@ -1480,6 +1486,10 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro
 		if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) {
 			continue
 		}
+		// 配额检查
+		if !s.isAccountSchedulableForQuota(acc) {
+			continue
+		}
 		// 窗口费用检查（非粘性会话路径）
 		if !s.isAccountSchedulableForWindowCost(ctx, acc, false) {
 			continue
@@ -2113,6 +2123,15 @@ func (s *GatewayService) withWindowCostPrefetch(ctx context.Context, accounts []
 	return context.WithValue(ctx, windowCostPrefetchContextKey, costs)
 }
 
+// isAccountSchedulableForQuota 检查 API Key 账号是否在配额限制内
+// 仅适用于配置了 quota_limit 的 apikey 类型账号
+func (s *GatewayService) isAccountSchedulableForQuota(account *Account) bool {
+	if account.Type != AccountTypeAPIKey {
+		return true
+	}
+	return !account.IsQuotaExceeded()
+}
+
 // isAccountSchedulableForWindowCost 检查账号是否可根据窗口费用进行调度
 // 仅适用于 Anthropic OAuth/SetupToken 账号
 // 返回 true 表示可调度，false 表示不可调度
@@ -2590,7 +2609,7 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context,
 						if clearSticky {
 							_ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash)
 						}
-						if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
+						if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
 							if s.debugModelRoutingEnabled() {
 								logger.LegacyPrintf("service.gateway", "[ModelRoutingDebug] legacy routed sticky hit: group_id=%v model=%s session=%s account=%d", derefGroupID(groupID), requestedModel, shortSessionHash(sessionHash), accountID)
 							}
@@ -2644,6 +2663,9 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context,
 			if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) {
 				continue
 			}
+			if !s.isAccountSchedulableForQuota(acc) {
+				continue
+			}
 			if !s.isAccountSchedulableForWindowCost(ctx, acc, false) {
 				continue
 			}
@@ -2700,7 +2722,7 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context,
 					if clearSticky {
 						_ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash)
 					}
-					if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
+					if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
 						return account, nil
 					}
 				}
@@ -2743,6 +2765,9 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context,
 		if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) {
 			continue
 		}
+		if !s.isAccountSchedulableForQuota(acc) {
+			continue
+		}
 		if !s.isAccountSchedulableForWindowCost(ctx, acc, false) {
 			continue
 		}
@@ -2818,7 +2843,7 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g
 						if clearSticky {
 							_ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash)
 						}
-						if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
+						if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
 							if account.Platform == nativePlatform || (account.Platform == PlatformAntigravity && account.IsMixedSchedulingEnabled()) {
 								if s.debugModelRoutingEnabled() {
 									logger.LegacyPrintf("service.gateway", "[ModelRoutingDebug] legacy mixed routed sticky hit: group_id=%v model=%s session=%s account=%d", derefGroupID(groupID), requestedModel, shortSessionHash(sessionHash), accountID)
@@ -2874,6 +2899,9 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g
 			if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) {
 				continue
 			}
+			if !s.isAccountSchedulableForQuota(acc) {
+				continue
+			}
 			if !s.isAccountSchedulableForWindowCost(ctx, acc, false) {
 				continue
 			}
@@ -2930,7 +2958,7 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g
 					if clearSticky {
 						_ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash)
 					}
-					if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
+					if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) {
 						if account.Platform == nativePlatform || (account.Platform == PlatformAntigravity && account.IsMixedSchedulingEnabled()) {
 							return account, nil
 						}
@@ -2975,6 +3003,9 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g
 		if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) {
 			continue
 		}
+		if !s.isAccountSchedulableForQuota(acc) {
+			continue
+		}
 		if !s.isAccountSchedulableForWindowCost(ctx, acc, false) {
 			continue
 		}
@@ -6578,6 +6609,13 @@ func (s *GatewayService) RecordUsage(ctx context.Context, input *RecordUsageInpu
 		s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost)
 	}
 
+	// 更新 API Key 账号配额用量
+	if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 {
+		if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil {
+			slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err)
+		}
+	}
+
 	// Schedule batch update for account last_used_at
 	s.deferredService.ScheduleLastUsedUpdate(account.ID)
 
@@ -6775,6 +6813,13 @@ func (s *GatewayService) RecordUsageWithLongContext(ctx context.Context, input *
 		s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost)
 	}
 
+	// 更新 API Key 账号配额用量
+	if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 {
+		if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil {
+			slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err)
+		}
+	}
+
 	// Schedule batch update for account last_used_at
 	s.deferredService.ScheduleLastUsedUpdate(account.ID)
 

backend/internal/service/gemini_multiplatform_test.go

@@ -176,6 +176,14 @@ func (m *mockAccountRepoForGemini) BulkUpdate(ctx context.Context, ids []int64,
 	return 0, nil
 }
 
+func (m *mockAccountRepoForGemini) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error {
+	return nil
+}
+
+func (m *mockAccountRepoForGemini) ResetQuotaUsed(ctx context.Context, id int64) error {
+	return nil
+}
+
 // Verify interface implementation
 var _ AccountRepository = (*mockAccountRepoForGemini)(nil)
 

backend/internal/service/openai_gateway_service.go

@@ -3502,6 +3502,13 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec
 		s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost)
 	}
 
+	// 更新 API Key 账号配额用量
+	if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 {
+		if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil {
+			logger.LegacyPrintf("service.openai_gateway", "increment account quota used failed: account_id=%d cost=%f error=%v", account.ID, cost.TotalCost, err)
+		}
+	}
+
 	// Schedule batch update for account last_used_at
 	s.deferredService.ScheduleLastUsedUpdate(account.ID)
 

frontend/src/api/admin/accounts.ts

@@ -240,6 +240,18 @@ export async function clearRateLimit(id: number): Promise<Account> {
   return data
 }
 
+/**
+ * Reset account quota usage
+ * @param id - Account ID
+ * @returns Updated account
+ */
+export async function resetAccountQuota(id: number): Promise<Account> {
+  const { data } = await apiClient.post<Account>(
+    `/admin/accounts/${id}/reset-quota`
+  )
+  return data
+}
+
 /**
  * Get temporary unschedulable status
  * @param id - Account ID
@@ -576,6 +588,7 @@ export const accountsAPI = {
   getTodayStats,
   getBatchTodayStats,
   clearRateLimit,
+  resetAccountQuota,
   getTempUnschedulableStatus,
   resetTempUnschedulable,
   setSchedulable,

frontend/src/components/account/AccountCapacityCell.vue

@@ -71,6 +71,24 @@
         <span class="text-[9px] opacity-60">{{ rpmStrategyTag }}</span>
       </span>
     </div>
+
+    <!-- API Key 账号配额限制 -->
+    <div v-if="showQuotaLimit" class="flex items-center gap-1">
+      <span
+        :class="[
+          'inline-flex items-center gap-1 rounded-md px-1.5 py-0.5 text-[10px] font-medium',
+          quotaClass
+        ]"
+        :title="quotaTooltip"
+      >
+        <svg class="h-2.5 w-2.5" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="2">
+          <path stroke-linecap="round" stroke-linejoin="round" d="M2.25 18.75a60.07 60.07 0 0115.797 2.101c.727.198 1.453-.342 1.453-1.096V18.75M3.75 4.5v.75A.75.75 0 013 6h-.75m0 0v-.375c0-.621.504-1.125 1.125-1.125H20.25M2.25 6v9m18-10.5v.75c0 .414.336.75.75.75h.75m-1.5-1.5h.375c.621 0 1.125.504 1.125 1.125v9.75c0 .621-.504 1.125-1.125 1.125h-.375m1.5-1.5H21a.75.75 0 00-.75.75v.75m0 0H3.75m0 0h-.375a1.125 1.125 0 01-1.125-1.125V15m1.5 1.5v-.75A.75.75 0 003 15h-.75M15 10.5a3 3 0 11-6 0 3 3 0 016 0zm3 0h.008v.008H18V10.5zm-12 0h.008v.008H6V10.5z" />
+        </svg>
+        <span class="font-mono">${{ formatCost(currentQuotaUsed) }}</span>
+        <span class="text-gray-400 dark:text-gray-500">/</span>
+        <span class="font-mono">${{ formatCost(account.quota_limit) }}</span>
+      </span>
+    </div>
   </div>
 </template>
 
@@ -286,6 +304,48 @@ const rpmTooltip = computed(() => {
   }
 })
 
+// 是否显示配额限制（仅 apikey 类型且设置了 quota_limit）
+const showQuotaLimit = computed(() => {
+  return (
+    props.account.type === 'apikey' &&
+    props.account.quota_limit !== undefined &&
+    props.account.quota_limit !== null &&
+    props.account.quota_limit > 0
+  )
+})
+
+// 当前已用配额
+const currentQuotaUsed = computed(() => props.account.quota_used ?? 0)
+
+// 配额状态样式
+const quotaClass = computed(() => {
+  if (!showQuotaLimit.value) return ''
+
+  const used = currentQuotaUsed.value
+  const limit = props.account.quota_limit || 0
+
+  if (used >= limit) {
+    return 'bg-red-100 text-red-700 dark:bg-red-900/30 dark:text-red-400'
+  }
+  if (used >= limit * 0.8) {
+    return 'bg-yellow-100 text-yellow-700 dark:bg-yellow-900/30 dark:text-yellow-400'
+  }
+  return 'bg-emerald-100 text-emerald-700 dark:bg-emerald-900/30 dark:text-emerald-400'
+})
+
+// 配额提示文字
+const quotaTooltip = computed(() => {
+  if (!showQuotaLimit.value) return ''
+
+  const used = currentQuotaUsed.value
+  const limit = props.account.quota_limit || 0
+
+  if (used >= limit) {
+    return t('admin.accounts.capacity.quota.exceeded')
+  }
+  return t('admin.accounts.capacity.quota.normal')
+})
+
 // 格式化费用显示
 const formatCost = (value: number | null | undefined) => {
   if (value === null || value === undefined) return '0'

frontend/src/components/account/EditAccountModal.vue

@@ -759,6 +759,26 @@
         </div>
       </div>
 
+      <!-- API Key 账号配额限制 -->
+      <div
+        v-if="account?.type === 'apikey'"
+        class="border-t border-gray-200 pt-4 dark:border-dark-600"
+      >
+        <label class="input-label">{{ t('admin.accounts.quotaLimit') }}</label>
+        <div class="flex items-center gap-2">
+          <span class="text-gray-500">$</span>
+          <input
+            v-model.number="editQuotaLimit"
+            type="number"
+            min="0"
+            step="0.01"
+            class="input"
+            :placeholder="t('admin.accounts.quotaLimitPlaceholder')"
+          />
+        </div>
+        <p class="input-hint">{{ t('admin.accounts.quotaLimitHint') }}</p>
+      </div>
+
       <!-- OpenAI OAuth Codex 官方客户端限制开关 -->
       <div
         v-if="account?.platform === 'openai' && account?.type === 'oauth'"
@@ -1386,6 +1406,7 @@ const openaiOAuthResponsesWebSocketV2Mode = ref<OpenAIWSMode>(OPENAI_WS_MODE_OFF
 const openaiAPIKeyResponsesWebSocketV2Mode = ref<OpenAIWSMode>(OPENAI_WS_MODE_OFF)
 const codexCLIOnlyEnabled = ref(false)
 const anthropicPassthroughEnabled = ref(false)
+const editQuotaLimit = ref<number | null>(null)
 const openAIWSModeOptions = computed(() => [
   { value: OPENAI_WS_MODE_OFF, label: t('admin.accounts.openai.wsModeOff') },
   // TODO: ctx_pool 选项暂时隐藏，待测试完成后恢复
@@ -1541,6 +1562,14 @@ watch(
         anthropicPassthroughEnabled.value = extra?.anthropic_passthrough === true
       }
 
+      // Load quota limit for apikey accounts
+      if (newAccount.type === 'apikey') {
+        const quotaVal = extra?.quota_limit as number | undefined
+        editQuotaLimit.value = (quotaVal && quotaVal > 0) ? quotaVal : null
+      } else {
+        editQuotaLimit.value = null
+      }
+
       // Load antigravity model mapping (Antigravity 只支持映射模式)
       if (newAccount.platform === 'antigravity') {
         const credentials = newAccount.credentials as Record<string, unknown> | undefined
@@ -2283,6 +2312,19 @@ const handleSubmit = async () => {
       updatePayload.extra = newExtra
     }
 
+    // For apikey accounts, handle quota_limit in extra
+    if (props.account.type === 'apikey') {
+      const currentExtra = (updatePayload.extra as Record<string, unknown>) ||
+        (props.account.extra as Record<string, unknown>) || {}
+      const newExtra: Record<string, unknown> = { ...currentExtra }
+      if (editQuotaLimit.value != null && editQuotaLimit.value > 0) {
+        newExtra.quota_limit = editQuotaLimit.value
+      } else {
+        delete newExtra.quota_limit
+      }
+      updatePayload.extra = newExtra
+    }
+
     const canContinue = await ensureAntigravityMixedChannelConfirmed(async () => {
       await submitUpdateAccount(accountID, updatePayload)
     })

frontend/src/components/admin/account/AccountActionMenu.vue

@@ -41,6 +41,10 @@
               <Icon name="clock" size="sm" />
               {{ t('admin.accounts.clearRateLimit') }}
             </button>
+            <button v-if="hasQuotaLimit" @click="$emit('reset-quota', account); $emit('close')" class="flex w-full items-center gap-2 px-4 py-2 text-sm text-teal-600 hover:bg-gray-100 dark:hover:bg-dark-700">
+              <Icon name="refresh" size="sm" />
+              {{ t('admin.accounts.resetQuota') }}
+            </button>
           </template>
         </div>
       </div>
@@ -55,7 +59,7 @@ import { Icon } from '@/components/icons'
 import type { Account } from '@/types'
 
 const props = defineProps<{ show: boolean; account: Account | null; position: { top: number; left: number } | null }>()
-const emit = defineEmits(['close', 'test', 'stats', 'schedule', 'reauth', 'refresh-token', 'reset-status', 'clear-rate-limit'])
+const emit = defineEmits(['close', 'test', 'stats', 'schedule', 'reauth', 'refresh-token', 'reset-status', 'clear-rate-limit', 'reset-quota'])
 const { t } = useI18n()
 const isRateLimited = computed(() => {
   if (props.account?.rate_limit_reset_at && new Date(props.account.rate_limit_reset_at) > new Date()) {
@@ -71,6 +75,12 @@ const isRateLimited = computed(() => {
   return false
 })
 const isOverloaded = computed(() => props.account?.overload_until && new Date(props.account.overload_until) > new Date())
+const hasQuotaLimit = computed(() => {
+  return props.account?.type === 'apikey' &&
+    props.account?.quota_limit !== undefined &&
+    props.account?.quota_limit !== null &&
+    props.account?.quota_limit > 0
+})
 
 const handleKeydown = (event: KeyboardEvent) => {
   if (event.key === 'Escape') emit('close')

frontend/src/i18n/locales/en.ts

@@ -1734,6 +1734,10 @@ export default {
           stickyExemptWarning: 'RPM limit (Sticky Exempt) - Approaching limit',
           stickyExemptOver: 'RPM limit (Sticky Exempt) - Over limit, sticky only'
         },
+        quota: {
+          exceeded: 'Quota exceeded, account paused',
+          normal: 'Quota normal'
+        },
       },
       tempUnschedulable: {
         title: 'Temp Unschedulable',
@@ -1779,6 +1783,10 @@ export default {
         }
       },
       clearRateLimit: 'Clear Rate Limit',
+      resetQuota: 'Reset Quota',
+      quotaLimit: 'Quota Limit',
+      quotaLimitPlaceholder: '0 means unlimited',
+      quotaLimitHint: 'Set max spending limit (USD). Account will be paused when reached. Changing limit won\'t reset usage.',
       testConnection: 'Test Connection',
       reAuthorize: 'Re-Authorize',
       refreshToken: 'Refresh Token',

frontend/src/i18n/locales/zh.ts

@@ -1784,8 +1784,16 @@ export default {
           stickyExemptWarning: 'RPM 限制 (粘性豁免) - 接近阈值',
           stickyExemptOver: 'RPM 限制 (粘性豁免) - 超限，仅粘性会话'
         },
+        quota: {
+          exceeded: '配额已用完，账号暂停调度',
+          normal: '配额正常'
+        },
       },
       clearRateLimit: '清除速率限制',
+      resetQuota: '重置配额',
+      quotaLimit: '配额限制',
+      quotaLimitPlaceholder: '0 表示不限制',
+      quotaLimitHint: '设置最大使用额度（美元），达到后账号暂停调度。修改限额不会重置已用额度。',
       testConnection: '测试连接',
       reAuthorize: '重新授权',
       refreshToken: '刷新令牌',

frontend/src/types/index.ts

@@ -705,6 +705,10 @@ export interface Account {
   cache_ttl_override_enabled?: boolean | null
   cache_ttl_override_target?: string | null
 
+  // API Key 账号配额限制
+  quota_limit?: number | null
+  quota_used?: number | null
+
   // 运行时状态（仅当启用对应限制时返回）
   current_window_cost?: number | null // 当前窗口费用
   active_sessions?: number | null // 当前活跃会话数

frontend/src/views/admin/AccountsView.vue

@@ -261,7 +261,7 @@
     <AccountTestModal :show="showTest" :account="testingAcc" @close="closeTestModal" />
     <AccountStatsModal :show="showStats" :account="statsAcc" @close="closeStatsModal" />
     <ScheduledTestsPanel :show="showSchedulePanel" :account-id="scheduleAcc?.id ?? null" :model-options="scheduleModelOptions" @close="closeSchedulePanel" />
-    <AccountActionMenu :show="menu.show" :account="menu.acc" :position="menu.pos" @close="menu.show = false" @test="handleTest" @stats="handleViewStats" @schedule="handleSchedule" @reauth="handleReAuth" @refresh-token="handleRefresh" @reset-status="handleResetStatus" @clear-rate-limit="handleClearRateLimit" />
+    <AccountActionMenu :show="menu.show" :account="menu.acc" :position="menu.pos" @close="menu.show = false" @test="handleTest" @stats="handleViewStats" @schedule="handleSchedule" @reauth="handleReAuth" @refresh-token="handleRefresh" @reset-status="handleResetStatus" @clear-rate-limit="handleClearRateLimit" @reset-quota="handleResetQuota" />
     <SyncFromCrsModal :show="showSync" @close="showSync = false" @synced="reload" />
     <ImportDataModal :show="showImportData" @close="showImportData = false" @imported="handleDataImported" />
     <BulkEditAccountModal :show="showBulkEdit" :account-ids="selIds" :selected-platforms="selPlatforms" :selected-types="selTypes" :proxies="proxies" :groups="groups" @close="showBulkEdit = false" @updated="handleBulkUpdated" />
@@ -1125,6 +1125,16 @@ const handleClearRateLimit = async (a: Account) => {
     console.error('Failed to clear rate limit:', error)
   }
 }
+const handleResetQuota = async (a: Account) => {
+  try {
+    const updated = await adminAPI.accounts.resetAccountQuota(a.id)
+    patchAccountInList(updated)
+    enterAutoRefreshSilentWindow()
+    appStore.showSuccess(t('common.success'))
+  } catch (error) {
+    console.error('Failed to reset quota:', error)
+  }
+}
 const handleDelete = (a: Account) => { deletingAcc.value = a; showDeleteDialog.value = true }
 const confirmDelete = async () => { if(!deletingAcc.value) return; try { await adminAPI.accounts.delete(deletingAcc.value.id); showDeleteDialog.value = false; deletingAcc.value = null; reload() } catch (error) { console.error('Failed to delete account:', error) } }
 const handleToggleSchedulable = async (a: Account) => {