feat(Sora): 完成Sora网关接入与媒体能力

新增 Sora 网关路由、账号调度与同步服务\n补充媒体代理与签名 URL、模型列表动态拉取\n完善计费配置、前端支持与相关测试
2026-01-31 20:22:22 +08:00
parent 99dc3b59bc
commit 618a614cbf
67 changed files with 4840 additions and 202 deletions
--- a/backend/internal/handler/admin/group_handler.go
+++ b/backend/internal/handler/admin/group_handler.go
@@ -27,7 +27,7 @@ func NewGroupHandler(adminService service.AdminService) *GroupHandler {
 type CreateGroupRequest struct {
 	Name             string   `json:"name" binding:"required"`
 	Description      string   `json:"description"`
-	Platform         string   `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity"`
+	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"`
@@ -38,6 +38,10 @@ type CreateGroupRequest struct {
 	ImagePrice1K    *float64 `json:"image_price_1k"`
 	ImagePrice2K    *float64 `json:"image_price_2k"`
 	ImagePrice4K    *float64 `json:"image_price_4k"`
+	SoraImagePrice360         *float64 `json:"sora_image_price_360"`
+	SoraImagePrice540         *float64 `json:"sora_image_price_540"`
+	SoraVideoPricePerRequest  *float64 `json:"sora_video_price_per_request"`
+	SoraVideoPricePerRequestHD *float64 `json:"sora_video_price_per_request_hd"`
 	ClaudeCodeOnly  bool     `json:"claude_code_only"`
 	FallbackGroupID *int64   `json:"fallback_group_id"`
 	// 模型路由配置（仅 anthropic 平台使用）
@@ -49,7 +53,7 @@ type CreateGroupRequest struct {
 type UpdateGroupRequest struct {
 	Name             string   `json:"name"`
 	Description      string   `json:"description"`
-	Platform         string   `json:"platform" binding:"omitempty,oneof=anthropic openai gemini antigravity"`
+	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"`
@@ -61,6 +65,10 @@ type UpdateGroupRequest struct {
 	ImagePrice1K    *float64 `json:"image_price_1k"`
 	ImagePrice2K    *float64 `json:"image_price_2k"`
 	ImagePrice4K    *float64 `json:"image_price_4k"`
+	SoraImagePrice360         *float64 `json:"sora_image_price_360"`
+	SoraImagePrice540         *float64 `json:"sora_image_price_540"`
+	SoraVideoPricePerRequest  *float64 `json:"sora_video_price_per_request"`
+	SoraVideoPricePerRequestHD *float64 `json:"sora_video_price_per_request_hd"`
 	ClaudeCodeOnly  *bool    `json:"claude_code_only"`
 	FallbackGroupID *int64   `json:"fallback_group_id"`
 	// 模型路由配置（仅 anthropic 平台使用）
@@ -167,6 +175,10 @@ func (h *GroupHandler) Create(c *gin.Context) {
 		ImagePrice1K:        req.ImagePrice1K,
 		ImagePrice2K:        req.ImagePrice2K,
 		ImagePrice4K:        req.ImagePrice4K,
+		SoraImagePrice360:   req.SoraImagePrice360,
+		SoraImagePrice540:   req.SoraImagePrice540,
+		SoraVideoPricePerRequest:  req.SoraVideoPricePerRequest,
+		SoraVideoPricePerRequestHD: req.SoraVideoPricePerRequestHD,
 		ClaudeCodeOnly:      req.ClaudeCodeOnly,
 		FallbackGroupID:     req.FallbackGroupID,
 		ModelRouting:        req.ModelRouting,
@@ -209,6 +221,10 @@ func (h *GroupHandler) Update(c *gin.Context) {
 		ImagePrice1K:        req.ImagePrice1K,
 		ImagePrice2K:        req.ImagePrice2K,
 		ImagePrice4K:        req.ImagePrice4K,
+		SoraImagePrice360:   req.SoraImagePrice360,
+		SoraImagePrice540:   req.SoraImagePrice540,
+		SoraVideoPricePerRequest:  req.SoraVideoPricePerRequest,
+		SoraVideoPricePerRequestHD: req.SoraVideoPricePerRequestHD,
 		ClaudeCodeOnly:      req.ClaudeCodeOnly,
 		FallbackGroupID:     req.FallbackGroupID,
 		ModelRouting:        req.ModelRouting,
--- a/backend/internal/handler/admin/model_handler.go
+++ b/backend/internal/handler/admin/model_handler.go
@@ -0,0 +1,55 @@
+package admin
+
+import (
+	"net/http"
+	"strings"
+
+	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+// ModelHandler handles admin model listing requests.
+type ModelHandler struct {
+	sora2apiService *service.Sora2APIService
+}
+
+// NewModelHandler creates a new ModelHandler.
+func NewModelHandler(sora2apiService *service.Sora2APIService) *ModelHandler {
+	return &ModelHandler{
+		sora2apiService: sora2apiService,
+	}
+}
+
+// List handles listing models for a specific platform
+// GET /api/v1/admin/models?platform=sora
+func (h *ModelHandler) List(c *gin.Context) {
+	platform := strings.TrimSpace(strings.ToLower(c.Query("platform")))
+	if platform == "" {
+		response.BadRequest(c, "platform is required")
+		return
+	}
+
+	switch platform {
+	case service.PlatformSora:
+		if h.sora2apiService == nil || !h.sora2apiService.Enabled() {
+			response.Error(c, http.StatusServiceUnavailable, "sora2api not configured")
+			return
+		}
+		models, err := h.sora2apiService.ListModels(c.Request.Context())
+		if err != nil {
+			response.Error(c, http.StatusServiceUnavailable, "failed to fetch sora models")
+			return
+		}
+		ids := make([]string, 0, len(models))
+		for _, m := range models {
+			if strings.TrimSpace(m.ID) != "" {
+				ids = append(ids, m.ID)
+			}
+		}
+		response.Success(c, ids)
+	default:
+		response.BadRequest(c, "unsupported platform")
+	}
+}
--- a/backend/internal/handler/admin/model_handler_test.go
+++ b/backend/internal/handler/admin/model_handler_test.go
@@ -0,0 +1,87 @@
+package admin
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Wei-Shaw/sub2api/internal/config"
+	"github.com/Wei-Shaw/sub2api/internal/pkg/response"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+func TestModelHandlerListSoraSuccess(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{"object":"list","data":[{"id":"m1"},{"id":"m2"}]}`))
+	}))
+	t.Cleanup(upstream.Close)
+
+	cfg := &config.Config{}
+	cfg.Sora2API.BaseURL = upstream.URL
+	cfg.Sora2API.APIKey = "test-key"
+	soraService := service.NewSora2APIService(cfg)
+
+	h := NewModelHandler(soraService)
+	router := gin.New()
+	router.GET("/admin/models", h.List)
+
+	req := httptest.NewRequest(http.MethodGet, "/admin/models?platform=sora", nil)
+	recorder := httptest.NewRecorder()
+	router.ServeHTTP(recorder, req)
+
+	if recorder.Code != http.StatusOK {
+		t.Fatalf("status=%d body=%s", recorder.Code, recorder.Body.String())
+	}
+	var resp response.Response
+	if err := json.Unmarshal(recorder.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("解析响应失败: %v", err)
+	}
+	if resp.Code != 0 {
+		t.Fatalf("响应 code=%d", resp.Code)
+	}
+	data, ok := resp.Data.([]any)
+	if !ok {
+		t.Fatalf("响应 data 类型错误")
+	}
+	if len(data) != 2 {
+		t.Fatalf("模型数量不符: %d", len(data))
+	}
+}
+
+func TestModelHandlerListSoraNotConfigured(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+
+	h := NewModelHandler(&service.Sora2APIService{})
+	router := gin.New()
+	router.GET("/admin/models", h.List)
+
+	req := httptest.NewRequest(http.MethodGet, "/admin/models?platform=sora", nil)
+	recorder := httptest.NewRecorder()
+	router.ServeHTTP(recorder, req)
+
+	if recorder.Code != http.StatusServiceUnavailable {
+		t.Fatalf("status=%d body=%s", recorder.Code, recorder.Body.String())
+	}
+}
+
+func TestModelHandlerListInvalidPlatform(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+
+	h := NewModelHandler(&service.Sora2APIService{})
+	router := gin.New()
+	router.GET("/admin/models", h.List)
+
+	req := httptest.NewRequest(http.MethodGet, "/admin/models?platform=unknown", nil)
+	recorder := httptest.NewRecorder()
+	router.ServeHTTP(recorder, req)
+
+	if recorder.Code != http.StatusBadRequest {
+		t.Fatalf("status=%d body=%s", recorder.Code, recorder.Body.String())
+	}
+}
--- a/backend/internal/handler/dto/mappers.go
+++ b/backend/internal/handler/dto/mappers.go
@@ -136,6 +136,10 @@ func groupFromServiceBase(g *service.Group) Group {
 		ImagePrice1K:     g.ImagePrice1K,
 		ImagePrice2K:     g.ImagePrice2K,
 		ImagePrice4K:     g.ImagePrice4K,
+		SoraImagePrice360:         g.SoraImagePrice360,
+		SoraImagePrice540:         g.SoraImagePrice540,
+		SoraVideoPricePerRequest:  g.SoraVideoPricePerRequest,
+		SoraVideoPricePerRequestHD: g.SoraVideoPricePerRequestHD,
 		ClaudeCodeOnly:   g.ClaudeCodeOnly,
 		FallbackGroupID:  g.FallbackGroupID,
 		CreatedAt:        g.CreatedAt,
@@ -379,6 +383,7 @@ func usageLogFromServiceUser(l *service.UsageLog) UsageLog {
 		FirstTokenMs:          l.FirstTokenMs,
 		ImageCount:            l.ImageCount,
 		ImageSize:             l.ImageSize,
+		MediaType:             l.MediaType,
 		UserAgent:             l.UserAgent,
 		CreatedAt:             l.CreatedAt,
 		User:                  UserFromServiceShallow(l.User),
--- a/backend/internal/handler/dto/types.go
+++ b/backend/internal/handler/dto/types.go
@@ -61,6 +61,12 @@ type Group struct {
 	ImagePrice2K *float64 `json:"image_price_2k"`
 	ImagePrice4K *float64 `json:"image_price_4k"`

+	// Sora 按次计费配置
+	SoraImagePrice360         *float64 `json:"sora_image_price_360"`
+	SoraImagePrice540         *float64 `json:"sora_image_price_540"`
+	SoraVideoPricePerRequest  *float64 `json:"sora_video_price_per_request"`
+	SoraVideoPricePerRequestHD *float64 `json:"sora_video_price_per_request_hd"`
+
 	// Claude Code 客户端限制
 	ClaudeCodeOnly  bool   `json:"claude_code_only"`
 	FallbackGroupID *int64 `json:"fallback_group_id"`
@@ -246,6 +252,7 @@ type UsageLog struct {
 	// 图片生成字段
 	ImageCount int     `json:"image_count"`
 	ImageSize  *string `json:"image_size"`
+	MediaType  *string `json:"media_type"`

 	// User-Agent
 	UserAgent *string `json:"user_agent"`
--- a/backend/internal/handler/gateway_handler.go
+++ b/backend/internal/handler/gateway_handler.go
@@ -29,6 +29,7 @@ type GatewayHandler struct {
 	geminiCompatService       *service.GeminiMessagesCompatService
 	antigravityGatewayService *service.AntigravityGatewayService
 	userService               *service.UserService
+	sora2apiService           *service.Sora2APIService
 	billingCacheService       *service.BillingCacheService
 	concurrencyHelper         *ConcurrencyHelper
 	maxAccountSwitches        int
@@ -41,6 +42,7 @@ func NewGatewayHandler(
 	geminiCompatService *service.GeminiMessagesCompatService,
 	antigravityGatewayService *service.AntigravityGatewayService,
 	userService *service.UserService,
+	sora2apiService *service.Sora2APIService,
 	concurrencyService *service.ConcurrencyService,
 	billingCacheService *service.BillingCacheService,
 	cfg *config.Config,
@@ -62,6 +64,7 @@ func NewGatewayHandler(
 		geminiCompatService:       geminiCompatService,
 		antigravityGatewayService: antigravityGatewayService,
 		userService:               userService,
+		sora2apiService:           sora2apiService,
 		billingCacheService:       billingCacheService,
 		concurrencyHelper:         NewConcurrencyHelper(concurrencyService, SSEPingFormatClaude, pingInterval),
 		maxAccountSwitches:        maxAccountSwitches,
@@ -478,6 +481,26 @@ func (h *GatewayHandler) Models(c *gin.Context) {
 		groupID = &apiKey.Group.ID
 		platform = apiKey.Group.Platform
 	}
+	if forcedPlatform, ok := middleware2.GetForcePlatformFromContext(c); ok && strings.TrimSpace(forcedPlatform) != "" {
+		platform = forcedPlatform
+	}
+
+	if platform == service.PlatformSora {
+		if h.sora2apiService == nil || !h.sora2apiService.Enabled() {
+			h.errorResponse(c, http.StatusServiceUnavailable, "api_error", "sora2api not configured")
+			return
+		}
+		models, err := h.sora2apiService.ListModels(c.Request.Context())
+		if err != nil {
+			h.errorResponse(c, http.StatusServiceUnavailable, "api_error", "Failed to fetch Sora models")
+			return
+		}
+		c.JSON(http.StatusOK, gin.H{
+			"object": "list",
+			"data":   models,
+		})
+		return
+	}

 	// Get available models from account configurations (without platform filter)
 	availableModels := h.gatewayService.GetAvailableModels(c.Request.Context(), groupID, "")
--- a/backend/internal/handler/handler.go
+++ b/backend/internal/handler/handler.go
@@ -23,6 +23,7 @@ type AdminHandlers struct {
 	Subscription     *admin.SubscriptionHandler
 	Usage            *admin.UsageHandler
 	UserAttribute    *admin.UserAttributeHandler
+	Model            *admin.ModelHandler
 }

 // Handlers contains all HTTP handlers
@@ -36,6 +37,7 @@ type Handlers struct {
 	Admin         *AdminHandlers
 	Gateway       *GatewayHandler
 	OpenAIGateway *OpenAIGatewayHandler
+	SoraGateway   *SoraGatewayHandler
 	Setting       *SettingHandler
 }

--- a/backend/internal/handler/sora_gateway_handler.go
+++ b/backend/internal/handler/sora_gateway_handler.go
@@ -0,0 +1,474 @@
+package handler
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/hex"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"log"
+	"net/http"
+	"path"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/Wei-Shaw/sub2api/internal/config"
+	"github.com/Wei-Shaw/sub2api/internal/pkg/ip"
+	middleware2 "github.com/Wei-Shaw/sub2api/internal/server/middleware"
+	"github.com/Wei-Shaw/sub2api/internal/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+// SoraGatewayHandler handles Sora chat completions requests
+type SoraGatewayHandler struct {
+	gatewayService      *service.GatewayService
+	soraGatewayService  *service.SoraGatewayService
+	billingCacheService *service.BillingCacheService
+	concurrencyHelper   *ConcurrencyHelper
+	maxAccountSwitches  int
+	streamMode          string
+	sora2apiBaseURL     string
+	soraMediaSigningKey string
+}
+
+// NewSoraGatewayHandler creates a new SoraGatewayHandler
+func NewSoraGatewayHandler(
+	gatewayService *service.GatewayService,
+	soraGatewayService *service.SoraGatewayService,
+	concurrencyService *service.ConcurrencyService,
+	billingCacheService *service.BillingCacheService,
+	cfg *config.Config,
+) *SoraGatewayHandler {
+	pingInterval := time.Duration(0)
+	maxAccountSwitches := 3
+	streamMode := "force"
+	signKey := ""
+	if cfg != nil {
+		pingInterval = time.Duration(cfg.Concurrency.PingInterval) * time.Second
+		if cfg.Gateway.MaxAccountSwitches > 0 {
+			maxAccountSwitches = cfg.Gateway.MaxAccountSwitches
+		}
+		if mode := strings.TrimSpace(cfg.Gateway.SoraStreamMode); mode != "" {
+			streamMode = mode
+		}
+		signKey = strings.TrimSpace(cfg.Gateway.SoraMediaSigningKey)
+	}
+	baseURL := ""
+	if cfg != nil {
+		baseURL = strings.TrimRight(strings.TrimSpace(cfg.Sora2API.BaseURL), "/")
+	}
+	return &SoraGatewayHandler{
+		gatewayService:      gatewayService,
+		soraGatewayService:  soraGatewayService,
+		billingCacheService: billingCacheService,
+		concurrencyHelper:   NewConcurrencyHelper(concurrencyService, SSEPingFormatComment, pingInterval),
+		maxAccountSwitches:  maxAccountSwitches,
+		streamMode:          strings.ToLower(streamMode),
+		sora2apiBaseURL:     baseURL,
+		soraMediaSigningKey: signKey,
+	}
+}
+
+// ChatCompletions handles Sora /v1/chat/completions endpoint
+func (h *SoraGatewayHandler) ChatCompletions(c *gin.Context) {
+	apiKey, ok := middleware2.GetAPIKeyFromContext(c)
+	if !ok {
+		h.errorResponse(c, http.StatusUnauthorized, "authentication_error", "Invalid API key")
+		return
+	}
+
+	subject, ok := middleware2.GetAuthSubjectFromContext(c)
+	if !ok {
+		h.errorResponse(c, http.StatusInternalServerError, "api_error", "User context not found")
+		return
+	}
+
+	body, err := io.ReadAll(c.Request.Body)
+	if err != nil {
+		if maxErr, ok := extractMaxBytesError(err); ok {
+			h.errorResponse(c, http.StatusRequestEntityTooLarge, "invalid_request_error", buildBodyTooLargeMessage(maxErr.Limit))
+			return
+		}
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "Failed to read request body")
+		return
+	}
+	if len(body) == 0 {
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "Request body is empty")
+		return
+	}
+
+	setOpsRequestContext(c, "", false, body)
+
+	var reqBody map[string]any
+	if err := json.Unmarshal(body, &reqBody); err != nil {
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "Failed to parse request body")
+		return
+	}
+
+	reqModel, _ := reqBody["model"].(string)
+	if reqModel == "" {
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "model is required")
+		return
+	}
+	reqMessages, _ := reqBody["messages"].([]any)
+	if len(reqMessages) == 0 {
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "messages is required")
+		return
+	}
+
+	clientStream, _ := reqBody["stream"].(bool)
+	if !clientStream {
+		if h.streamMode == "error" {
+			h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "Sora requires stream=true")
+			return
+		}
+		reqBody["stream"] = true
+		updated, err := json.Marshal(reqBody)
+		if err != nil {
+			h.errorResponse(c, http.StatusInternalServerError, "api_error", "Failed to process request")
+			return
+		}
+		body = updated
+	}
+
+	setOpsRequestContext(c, reqModel, clientStream, body)
+
+	platform := ""
+	if forced, ok := middleware2.GetForcePlatformFromContext(c); ok {
+		platform = forced
+	} else if apiKey.Group != nil {
+		platform = apiKey.Group.Platform
+	}
+	if platform != service.PlatformSora {
+		h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "This endpoint only supports Sora platform")
+		return
+	}
+
+	streamStarted := false
+	subscription, _ := middleware2.GetSubscriptionFromContext(c)
+
+	maxWait := service.CalculateMaxWait(subject.Concurrency)
+	canWait, err := h.concurrencyHelper.IncrementWaitCount(c.Request.Context(), subject.UserID, maxWait)
+	waitCounted := false
+	if err != nil {
+		log.Printf("Increment wait count failed: %v", err)
+	} else if !canWait {
+		h.errorResponse(c, http.StatusTooManyRequests, "rate_limit_error", "Too many pending requests, please retry later")
+		return
+	}
+	if err == nil && canWait {
+		waitCounted = true
+	}
+	defer func() {
+		if waitCounted {
+			h.concurrencyHelper.DecrementWaitCount(c.Request.Context(), subject.UserID)
+		}
+	}()
+
+	userReleaseFunc, err := h.concurrencyHelper.AcquireUserSlotWithWait(c, subject.UserID, subject.Concurrency, clientStream, &streamStarted)
+	if err != nil {
+		log.Printf("User concurrency acquire failed: %v", err)
+		h.handleConcurrencyError(c, err, "user", streamStarted)
+		return
+	}
+	if waitCounted {
+		h.concurrencyHelper.DecrementWaitCount(c.Request.Context(), subject.UserID)
+		waitCounted = false
+	}
+	userReleaseFunc = wrapReleaseOnDone(c.Request.Context(), userReleaseFunc)
+	if userReleaseFunc != nil {
+		defer userReleaseFunc()
+	}
+
+	if err := h.billingCacheService.CheckBillingEligibility(c.Request.Context(), apiKey.User, apiKey, apiKey.Group, subscription); err != nil {
+		log.Printf("Billing eligibility check failed after wait: %v", err)
+		status, code, message := billingErrorDetails(err)
+		h.handleStreamingAwareError(c, status, code, message, streamStarted)
+		return
+	}
+
+	sessionHash := generateOpenAISessionHash(c, reqBody)
+
+	maxAccountSwitches := h.maxAccountSwitches
+	switchCount := 0
+	failedAccountIDs := make(map[int64]struct{})
+	lastFailoverStatus := 0
+
+	for {
+		selection, err := h.gatewayService.SelectAccountWithLoadAwareness(c.Request.Context(), apiKey.GroupID, sessionHash, reqModel, failedAccountIDs, "")
+		if err != nil {
+			log.Printf("[Sora Handler] SelectAccount failed: %v", err)
+			if len(failedAccountIDs) == 0 {
+				h.handleStreamingAwareError(c, http.StatusServiceUnavailable, "api_error", "No available accounts: "+err.Error(), streamStarted)
+				return
+			}
+			h.handleFailoverExhausted(c, lastFailoverStatus, streamStarted)
+			return
+		}
+		account := selection.Account
+		setOpsSelectedAccount(c, account.ID)
+
+		accountReleaseFunc := selection.ReleaseFunc
+		if !selection.Acquired {
+			if selection.WaitPlan == nil {
+				h.handleStreamingAwareError(c, http.StatusServiceUnavailable, "api_error", "No available accounts", streamStarted)
+				return
+			}
+			accountWaitCounted := false
+			canWait, err := h.concurrencyHelper.IncrementAccountWaitCount(c.Request.Context(), account.ID, selection.WaitPlan.MaxWaiting)
+			if err != nil {
+				log.Printf("Increment account wait count failed: %v", err)
+			} else if !canWait {
+				log.Printf("Account wait queue full: account=%d", account.ID)
+				h.handleStreamingAwareError(c, http.StatusTooManyRequests, "rate_limit_error", "Too many pending requests, please retry later", streamStarted)
+				return
+			}
+			if err == nil && canWait {
+				accountWaitCounted = true
+			}
+			defer func() {
+				if accountWaitCounted {
+					h.concurrencyHelper.DecrementAccountWaitCount(c.Request.Context(), account.ID)
+				}
+			}()
+
+			accountReleaseFunc, err = h.concurrencyHelper.AcquireAccountSlotWithWaitTimeout(
+				c,
+				account.ID,
+				selection.WaitPlan.MaxConcurrency,
+				selection.WaitPlan.Timeout,
+				clientStream,
+				&streamStarted,
+			)
+			if err != nil {
+				log.Printf("Account concurrency acquire failed: %v", err)
+				h.handleConcurrencyError(c, err, "account", streamStarted)
+				return
+			}
+			if accountWaitCounted {
+				h.concurrencyHelper.DecrementAccountWaitCount(c.Request.Context(), account.ID)
+				accountWaitCounted = false
+			}
+		}
+		accountReleaseFunc = wrapReleaseOnDone(c.Request.Context(), accountReleaseFunc)
+
+		result, err := h.soraGatewayService.Forward(c.Request.Context(), c, account, body, clientStream)
+		if accountReleaseFunc != nil {
+			accountReleaseFunc()
+		}
+		if err != nil {
+			var failoverErr *service.UpstreamFailoverError
+			if errors.As(err, &failoverErr) {
+				failedAccountIDs[account.ID] = struct{}{}
+				if switchCount >= maxAccountSwitches {
+					lastFailoverStatus = failoverErr.StatusCode
+					h.handleFailoverExhausted(c, lastFailoverStatus, streamStarted)
+					return
+				}
+				lastFailoverStatus = failoverErr.StatusCode
+				switchCount++
+				log.Printf("Account %d: upstream error %d, switching account %d/%d", account.ID, failoverErr.StatusCode, switchCount, maxAccountSwitches)
+				continue
+			}
+			log.Printf("Account %d: Forward request failed: %v", account.ID, err)
+			return
+		}
+
+		userAgent := c.GetHeader("User-Agent")
+		clientIP := ip.GetClientIP(c)
+
+		go func(result *service.ForwardResult, usedAccount *service.Account, ua, ip string) {
+			ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+			defer cancel()
+			if err := h.gatewayService.RecordUsage(ctx, &service.RecordUsageInput{
+				Result:       result,
+				APIKey:       apiKey,
+				User:         apiKey.User,
+				Account:      usedAccount,
+				Subscription: subscription,
+				UserAgent:    ua,
+				IPAddress:    ip,
+			}); err != nil {
+				log.Printf("Record usage failed: %v", err)
+			}
+		}(result, account, userAgent, clientIP)
+		return
+	}
+}
+
+func generateOpenAISessionHash(c *gin.Context, reqBody map[string]any) string {
+	if c == nil {
+		return ""
+	}
+	sessionID := strings.TrimSpace(c.GetHeader("session_id"))
+	if sessionID == "" {
+		sessionID = strings.TrimSpace(c.GetHeader("conversation_id"))
+	}
+	if sessionID == "" && reqBody != nil {
+		if v, ok := reqBody["prompt_cache_key"].(string); ok {
+			sessionID = strings.TrimSpace(v)
+		}
+	}
+	if sessionID == "" {
+		return ""
+	}
+	hash := sha256.Sum256([]byte(sessionID))
+	return hex.EncodeToString(hash[:])
+}
+
+func (h *SoraGatewayHandler) handleConcurrencyError(c *gin.Context, err error, slotType string, streamStarted bool) {
+	h.handleStreamingAwareError(c, http.StatusTooManyRequests, "rate_limit_error",
+		fmt.Sprintf("Concurrency limit exceeded for %s, please retry later", slotType), streamStarted)
+}
+
+func (h *SoraGatewayHandler) handleFailoverExhausted(c *gin.Context, statusCode int, streamStarted bool) {
+	status, errType, errMsg := h.mapUpstreamError(statusCode)
+	h.handleStreamingAwareError(c, status, errType, errMsg, streamStarted)
+}
+
+func (h *SoraGatewayHandler) mapUpstreamError(statusCode int) (int, string, string) {
+	switch statusCode {
+	case 401:
+		return http.StatusBadGateway, "upstream_error", "Upstream authentication failed, please contact administrator"
+	case 403:
+		return http.StatusBadGateway, "upstream_error", "Upstream access forbidden, please contact administrator"
+	case 429:
+		return http.StatusTooManyRequests, "rate_limit_error", "Upstream rate limit exceeded, please retry later"
+	case 529:
+		return http.StatusServiceUnavailable, "upstream_error", "Upstream service overloaded, please retry later"
+	case 500, 502, 503, 504:
+		return http.StatusBadGateway, "upstream_error", "Upstream service temporarily unavailable"
+	default:
+		return http.StatusBadGateway, "upstream_error", "Upstream request failed"
+	}
+}
+
+func (h *SoraGatewayHandler) handleStreamingAwareError(c *gin.Context, status int, errType, message string, streamStarted bool) {
+	if streamStarted {
+		flusher, ok := c.Writer.(http.Flusher)
+		if ok {
+			errorEvent := fmt.Sprintf(`event: error`+"\n"+`data: {"error": {"type": "%s", "message": "%s"}}`+"\n\n", errType, message)
+			if _, err := fmt.Fprint(c.Writer, errorEvent); err != nil {
+				_ = c.Error(err)
+			}
+			flusher.Flush()
+		}
+		return
+	}
+	h.errorResponse(c, status, errType, message)
+}
+
+func (h *SoraGatewayHandler) errorResponse(c *gin.Context, status int, errType, message string) {
+	c.JSON(status, gin.H{
+		"error": gin.H{
+			"type":    errType,
+			"message": message,
+		},
+	})
+}
+
+// MediaProxy proxies /tmp or /static media files from sora2api
+func (h *SoraGatewayHandler) MediaProxy(c *gin.Context) {
+	h.proxySoraMedia(c, false)
+}
+
+// MediaProxySigned proxies /tmp or /static media files with signature verification
+func (h *SoraGatewayHandler) MediaProxySigned(c *gin.Context) {
+	h.proxySoraMedia(c, true)
+}
+
+func (h *SoraGatewayHandler) proxySoraMedia(c *gin.Context, requireSignature bool) {
+	if h.sora2apiBaseURL == "" {
+		c.JSON(http.StatusServiceUnavailable, gin.H{
+			"error": gin.H{
+				"type":    "api_error",
+				"message": "sora2api 未配置",
+			},
+		})
+		return
+	}
+
+	rawPath := c.Param("filepath")
+	if rawPath == "" {
+		c.Status(http.StatusNotFound)
+		return
+	}
+	cleaned := path.Clean(rawPath)
+	if !strings.HasPrefix(cleaned, "/tmp/") && !strings.HasPrefix(cleaned, "/static/") {
+		c.Status(http.StatusNotFound)
+		return
+	}
+
+	query := c.Request.URL.Query()
+	if requireSignature {
+		if h.soraMediaSigningKey == "" {
+			c.JSON(http.StatusServiceUnavailable, gin.H{
+				"error": gin.H{
+					"type":    "api_error",
+					"message": "Sora 媒体签名未配置",
+				},
+			})
+			return
+		}
+		expiresStr := strings.TrimSpace(query.Get("expires"))
+		signature := strings.TrimSpace(query.Get("sig"))
+		expires, err := strconv.ParseInt(expiresStr, 10, 64)
+		if err != nil || expires <= time.Now().Unix() {
+			c.JSON(http.StatusUnauthorized, gin.H{
+				"error": gin.H{
+					"type":    "authentication_error",
+					"message": "Sora 媒体签名已过期",
+				},
+			})
+			return
+		}
+		query.Del("sig")
+		query.Del("expires")
+		signingQuery := query.Encode()
+		if !service.VerifySoraMediaURL(cleaned, signingQuery, expires, signature, h.soraMediaSigningKey) {
+			c.JSON(http.StatusUnauthorized, gin.H{
+				"error": gin.H{
+					"type":    "authentication_error",
+					"message": "Sora 媒体签名无效",
+				},
+			})
+			return
+		}
+	}
+
+	targetURL := h.sora2apiBaseURL + cleaned
+	if rawQuery := query.Encode(); rawQuery != "" {
+		targetURL += "?" + rawQuery
+	}
+
+	req, err := http.NewRequestWithContext(c.Request.Context(), c.Request.Method, targetURL, nil)
+	if err != nil {
+		c.Status(http.StatusBadGateway)
+		return
+	}
+	copyHeaders := []string{"Range", "If-Range", "If-Modified-Since", "If-None-Match", "Accept", "User-Agent"}
+	for _, key := range copyHeaders {
+		if val := c.GetHeader(key); val != "" {
+			req.Header.Set(key, val)
+		}
+	}
+
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		c.Status(http.StatusBadGateway)
+		return
+	}
+	defer func() { _ = resp.Body.Close() }()
+
+	for _, key := range []string{"Content-Type", "Content-Length", "Accept-Ranges", "Content-Range", "Cache-Control", "Last-Modified", "ETag"} {
+		if val := resp.Header.Get(key); val != "" {
+			c.Header(key, val)
+		}
+	}
+	c.Status(resp.StatusCode)
+	_, _ = io.Copy(c.Writer, resp.Body)
+}
--- a/backend/internal/handler/wire.go
+++ b/backend/internal/handler/wire.go
@@ -26,6 +26,7 @@ func ProvideAdminHandlers(
 	subscriptionHandler *admin.SubscriptionHandler,
 	usageHandler *admin.UsageHandler,
 	userAttributeHandler *admin.UserAttributeHandler,
+	modelHandler *admin.ModelHandler,
 ) *AdminHandlers {
 	return &AdminHandlers{
 		Dashboard:        dashboardHandler,
@@ -45,6 +46,7 @@ func ProvideAdminHandlers(
 		Subscription:     subscriptionHandler,
 		Usage:            usageHandler,
 		UserAttribute:    userAttributeHandler,
+		Model:            modelHandler,
 	}
 }

@@ -69,6 +71,7 @@ func ProvideHandlers(
 	adminHandlers *AdminHandlers,
 	gatewayHandler *GatewayHandler,
 	openaiGatewayHandler *OpenAIGatewayHandler,
+	soraGatewayHandler *SoraGatewayHandler,
 	settingHandler *SettingHandler,
 ) *Handlers {
 	return &Handlers{
@@ -81,6 +84,7 @@ func ProvideHandlers(
 		Admin:         adminHandlers,
 		Gateway:       gatewayHandler,
 		OpenAIGateway: openaiGatewayHandler,
+		SoraGateway:   soraGatewayHandler,
 		Setting:       settingHandler,
 	}
 }
@@ -96,6 +100,7 @@ var ProviderSet = wire.NewSet(
 	NewSubscriptionHandler,
 	NewGatewayHandler,
 	NewOpenAIGatewayHandler,
+	NewSoraGatewayHandler,
 	ProvideSettingHandler,

 	// Admin handlers
@@ -116,6 +121,7 @@ var ProviderSet = wire.NewSet(
 	admin.NewSubscriptionHandler,
 	admin.NewUsageHandler,
 	admin.NewUserAttributeHandler,
+	admin.NewModelHandler,

 	// AdminHandlers and Handlers constructors
 	ProvideAdminHandlers,