feat(channel): add support for Vertex AI key type configuration in settings

fix: UI 未对齐问题

controller/channel-test.go

@@ -235,7 +235,7 @@ func testChannel(channel *model.Channel, testModel string) testResult {
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			err := service.RelayErrorHandler(httpResp, true)
+			err := service.RelayErrorHandler(c.Request.Context(), httpResp, true)
 			return testResult{
 				context:     c,
 				localErr:    err,

controller/channel.go

@@ -6,6 +6,7 @@ import (
 	"net/http"
 	"one-api/common"
 	"one-api/constant"
+	"one-api/dto"
 	"one-api/model"
 	"strconv"
 	"strings"
@@ -560,7 +561,7 @@ func AddChannel(c *gin.Context) {
 	case "multi_to_single":
 		addChannelRequest.Channel.ChannelInfo.IsMultiKey = true
 		addChannelRequest.Channel.ChannelInfo.MultiKeyMode = addChannelRequest.MultiKeyMode
-		if addChannelRequest.Channel.Type == constant.ChannelTypeVertexAi {
+		if addChannelRequest.Channel.Type == constant.ChannelTypeVertexAi && addChannelRequest.Channel.GetOtherSettings().VertexKeyType != dto.VertexKeyTypeAPIKey {
 			array, err := getVertexArrayKeys(addChannelRequest.Channel.Key)
 			if err != nil {
 				c.JSON(http.StatusOK, gin.H{
@@ -585,7 +586,7 @@ func AddChannel(c *gin.Context) {
 		}
 		keys = []string{addChannelRequest.Channel.Key}
 	case "batch":
-		if addChannelRequest.Channel.Type == constant.ChannelTypeVertexAi {
+		if addChannelRequest.Channel.Type == constant.ChannelTypeVertexAi && addChannelRequest.Channel.GetOtherSettings().VertexKeyType != dto.VertexKeyTypeAPIKey {
 			// multi json
 			keys, err = getVertexArrayKeys(addChannelRequest.Channel.Key)
 			if err != nil {
@@ -840,7 +841,7 @@ func UpdateChannel(c *gin.Context) {
 				}
 
 				// 处理 Vertex AI 的特殊情况
-				if channel.Type == constant.ChannelTypeVertexAi {
+				if channel.Type == constant.ChannelTypeVertexAi && channel.GetOtherSettings().VertexKeyType != dto.VertexKeyTypeAPIKey {
 					// 尝试解析新密钥为JSON数组
 					if strings.HasPrefix(strings.TrimSpace(channel.Key), "[") {
 						array, err := getVertexArrayKeys(channel.Key)

controller/relay.go

@@ -139,15 +139,15 @@ func Relay(c *gin.Context, relayFormat types.RelayFormat) {
 
 	// common.SetContextKey(c, constant.ContextKeyTokenCountMeta, meta)
 
-	preConsumedQuota, newAPIError := service.PreConsumeQuota(c, priceData.ShouldPreConsumedQuota, relayInfo)
+	newAPIError = service.PreConsumeQuota(c, priceData.ShouldPreConsumedQuota, relayInfo)
 	if newAPIError != nil {
 		return
 	}
 
 	defer func() {
 		// Only return quota if downstream failed and quota was actually pre-consumed
-		if newAPIError != nil && preConsumedQuota != 0 {
-			service.ReturnPreConsumedQuota(c, relayInfo, preConsumedQuota)
+		if newAPIError != nil && relayInfo.FinalPreConsumedQuota != 0 {
+			service.ReturnPreConsumedQuota(c, relayInfo)
 		}
 	}()
 
@@ -277,14 +277,13 @@ func shouldRetry(c *gin.Context, openaiErr *types.NewAPIError, retryTimes int) b
 
 func processChannelError(c *gin.Context, channelError types.ChannelError, err *types.NewAPIError) {
 	logger.LogError(c, fmt.Sprintf("relay error (channel #%d, status code: %d): %s", channelError.ChannelId, err.StatusCode, err.Error()))
-
-	gopool.Go(func() {
-		// 不要使用context获取渠道信息，异步处理时可能会出现渠道信息不一致的情况
-		// do not use context to get channel info, there may be inconsistent channel info when processing asynchronously
-		if service.ShouldDisableChannel(channelError.ChannelId, err) && channelError.AutoBan {
+	// 不要使用context获取渠道信息，异步处理时可能会出现渠道信息不一致的情况
+	// do not use context to get channel info, there may be inconsistent channel info when processing asynchronously
+	if service.ShouldDisableChannel(channelError.ChannelId, err) && channelError.AutoBan {
+		gopool.Go(func() {
 			service.DisableChannel(channelError, err.Error())
-		}
-	})
+		})
+	}
 
 	if constant.ErrorLogEnabled && types.IsRecordErrorLog(err) {
 		// 保存错误日志到mysql中

dto/channel_settings.go

@@ -9,6 +9,14 @@ type ChannelSettings struct {
 	SystemPromptOverride   bool   `json:"system_prompt_override,omitempty"`
 }
 
+type VertexKeyType string
+
+const (
+	VertexKeyTypeJSON   VertexKeyType = "json"
+	VertexKeyTypeAPIKey VertexKeyType = "api_key"
+)
+
 type ChannelOtherSettings struct {
-	AzureResponsesVersion string `json:"azure_responses_version,omitempty"`
+	AzureResponsesVersion string        `json:"azure_responses_version,omitempty"`
+	VertexKeyType         VertexKeyType `json:"vertex_key_type,omitempty"` // "json" or "api_key"
 }

dto/openai_image.go

@@ -59,6 +59,31 @@ func (i *ImageRequest) UnmarshalJSON(data []byte) error {
 	return nil
 }
 
+// 序列化时需要重新把字段平铺
+func (r ImageRequest) MarshalJSON() ([]byte, error) {
+	// 将已定义字段转为 map
+	type Alias ImageRequest
+	alias := Alias(r)
+	base, err := common.Marshal(alias)
+	if err != nil {
+		return nil, err
+	}
+
+	var baseMap map[string]json.RawMessage
+	if err := common.Unmarshal(base, &baseMap); err != nil {
+		return nil, err
+	}
+
+	// 合并 ExtraFields
+	for k, v := range r.Extra {
+		if _, exists := baseMap[k]; !exists {
+			baseMap[k] = v
+		}
+	}
+
+	return json.Marshal(baseMap)
+}
+
 func GetJSONFieldNames(t reflect.Type) map[string]struct{} {
 	fields := make(map[string]struct{})
 	for i := 0; i < t.NumField(); i++ {

model/channel.go

@@ -42,7 +42,6 @@ type Channel struct {
 	Priority          *int64  `json:"priority" gorm:"bigint;default:0"`
 	AutoBan           *int    `json:"auto_ban" gorm:"default:1"`
 	OtherInfo         string  `json:"other_info"`
-	OtherSettings     string  `json:"settings" gorm:"column:settings"` // 其他设置
 	Tag               *string `json:"tag" gorm:"index"`
 	Setting           *string `json:"setting" gorm:"type:text"` // 渠道额外设置
 	ParamOverride     *string `json:"param_override" gorm:"type:text"`
@@ -51,6 +50,8 @@ type Channel struct {
 	// add after v0.8.5
 	ChannelInfo ChannelInfo `json:"channel_info" gorm:"type:json"`
 
+	OtherSettings string `json:"settings" gorm:"column:settings"` // 其他设置，存储azure版本等不需要检索的信息，详见dto.ChannelOtherSettings
+
 	// cache info
 	Keys []string `json:"-" gorm:"-"`
 }

relay/audio_handler.go

@@ -53,7 +53,7 @@ func AudioHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *type
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError

relay/channel/api_request.go

@@ -264,9 +264,8 @@ func doRequest(c *gin.Context, req *http.Request, info *common.RelayInfo) (*http
 	}
 
 	resp, err := client.Do(req)
-
 	if err != nil {
-		return nil, err
+		return nil, types.NewError(err, types.ErrorCodeDoRequestFailed, types.ErrOptionWithHideErrMsg("upstream error: do request failed"))
 	}
 	if resp == nil {
 		return nil, errors.New("resp is nil")

relay/channel/aws/adaptor.go

@@ -60,7 +60,16 @@ func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayIn
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
+	// 检查是否为Nova模型
+	if isNovaModel(request.Model) {
+		novaReq := convertToNovaRequest(request)
+		c.Set("request_model", request.Model)
+		c.Set("converted_request", novaReq)
+		c.Set("is_nova_model", true)
+		return novaReq, nil
+	}
 
+	// 原有的Claude模型处理逻辑
 	var claudeReq *dto.ClaudeRequest
 	var err error
 	claudeReq, err = claude.RequestOpenAI2ClaudeMessage(c, *request)
@@ -69,6 +78,7 @@ func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayIn
 	}
 	c.Set("request_model", claudeReq.Model)
 	c.Set("converted_request", claudeReq)
+	c.Set("is_nova_model", false)
 	return claudeReq, err
 }
 

relay/channel/aws/constants.go

@@ -1,5 +1,7 @@
 package aws
 
+import "strings"
+
 var awsModelIDMap = map[string]string{
 	"claude-instant-1.2":         "anthropic.claude-instant-v1",
 	"claude-2.0":                 "anthropic.claude-v2",
@@ -14,6 +16,11 @@ var awsModelIDMap = map[string]string{
 	"claude-sonnet-4-20250514":   "anthropic.claude-sonnet-4-20250514-v1:0",
 	"claude-opus-4-20250514":     "anthropic.claude-opus-4-20250514-v1:0",
 	"claude-opus-4-1-20250805":   "anthropic.claude-opus-4-1-20250805-v1:0",
+	// Nova models
+	"nova-micro-v1:0":   "amazon.nova-micro-v1:0",
+	"nova-lite-v1:0":    "amazon.nova-lite-v1:0",
+	"nova-pro-v1:0":     "amazon.nova-pro-v1:0",
+	"nova-premier-v1:0": "amazon.nova-premier-v1:0",
 }
 
 var awsModelCanCrossRegionMap = map[string]map[string]bool{
@@ -58,7 +65,27 @@ var awsModelCanCrossRegionMap = map[string]map[string]bool{
 	"anthropic.claude-opus-4-1-20250805-v1:0": {
 		"us": true,
 	},
-}
+	// Nova models - all support three major regions
+	"amazon.nova-micro-v1:0": {
+		"us":   true,
+		"eu":   true,
+		"apac": true,
+	},
+	"amazon.nova-lite-v1:0": {
+		"us":   true,
+		"eu":   true,
+		"apac": true,
+	},
+	"amazon.nova-pro-v1:0": {
+		"us":   true,
+		"eu":   true,
+		"apac": true,
+	},
+	"amazon.nova-premier-v1:0": {
+		"us":   true,
+		"eu":   true,
+		"apac": true,
+	}}
 
 var awsRegionCrossModelPrefixMap = map[string]string{
 	"us": "us",
@@ -67,3 +94,8 @@ var awsRegionCrossModelPrefixMap = map[string]string{
 }
 
 var ChannelName = "aws"
+
+// 判断是否为Nova模型
+func isNovaModel(modelId string) bool {
+	return strings.HasPrefix(modelId, "nova-")
+}

relay/channel/aws/dto.go

@@ -34,3 +34,92 @@ func copyRequest(req *dto.ClaudeRequest) *AwsClaudeRequest {
 		Thinking:         req.Thinking,
 	}
 }
+
+// NovaMessage Nova模型使用messages-v1格式
+type NovaMessage struct {
+	Role    string        `json:"role"`
+	Content []NovaContent `json:"content"`
+}
+
+type NovaContent struct {
+	Text string `json:"text"`
+}
+
+type NovaRequest struct {
+	SchemaVersion   string               `json:"schemaVersion"`             // 请求版本，例如 "1.0"
+	Messages        []NovaMessage        `json:"messages"`                  // 对话消息列表
+	InferenceConfig *NovaInferenceConfig `json:"inferenceConfig,omitempty"` // 推理配置，可选
+}
+
+type NovaInferenceConfig struct {
+	MaxTokens     int      `json:"maxTokens,omitempty"`     // 最大生成的 token 数
+	Temperature   float64  `json:"temperature,omitempty"`   // 随机性 (默认 0.7, 范围 0-1)
+	TopP          float64  `json:"topP,omitempty"`          // nucleus sampling (默认 0.9, 范围 0-1)
+	TopK          int      `json:"topK,omitempty"`          // 限制候选 token 数 (默认 50, 范围 0-128)
+	StopSequences []string `json:"stopSequences,omitempty"` // 停止生成的序列
+}
+
+// 转换OpenAI请求为Nova格式
+func convertToNovaRequest(req *dto.GeneralOpenAIRequest) *NovaRequest {
+	novaMessages := make([]NovaMessage, len(req.Messages))
+	for i, msg := range req.Messages {
+		novaMessages[i] = NovaMessage{
+			Role:    msg.Role,
+			Content: []NovaContent{{Text: msg.StringContent()}},
+		}
+	}
+
+	novaReq := &NovaRequest{
+		SchemaVersion: "messages-v1",
+		Messages:      novaMessages,
+	}
+
+	// 设置推理配置
+	if req.MaxTokens != 0 || (req.Temperature != nil && *req.Temperature != 0) || req.TopP != 0 || req.TopK != 0 || req.Stop != nil {
+		novaReq.InferenceConfig = &NovaInferenceConfig{}
+		if req.MaxTokens != 0 {
+			novaReq.InferenceConfig.MaxTokens = int(req.MaxTokens)
+		}
+		if req.Temperature != nil && *req.Temperature != 0 {
+			novaReq.InferenceConfig.Temperature = *req.Temperature
+		}
+		if req.TopP != 0 {
+			novaReq.InferenceConfig.TopP = req.TopP
+		}
+		if req.TopK != 0 {
+			novaReq.InferenceConfig.TopK = req.TopK
+		}
+		if req.Stop != nil {
+			if stopSequences := parseStopSequences(req.Stop); len(stopSequences) > 0 {
+				novaReq.InferenceConfig.StopSequences = stopSequences
+			}
+		}
+	}
+
+	return novaReq
+}
+
+// parseStopSequences 解析停止序列，支持字符串或字符串数组
+func parseStopSequences(stop any) []string {
+	if stop == nil {
+		return nil
+	}
+
+	switch v := stop.(type) {
+	case string:
+		if v != "" {
+			return []string{v}
+		}
+	case []string:
+		return v
+	case []interface{}:
+		var sequences []string
+		for _, item := range v {
+			if str, ok := item.(string); ok && str != "" {
+				sequences = append(sequences, str)
+			}
+		}
+		return sequences
+	}
+	return nil
+}

relay/channel/aws/relay-aws.go

@@ -1,6 +1,7 @@
 package aws
 
 import (
+	"encoding/json"
 	"fmt"
 	"net/http"
 	"one-api/common"
@@ -93,7 +94,19 @@ func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*
 	}
 
 	awsModelId := awsModelID(c.GetString("request_model"))
+	// 检查是否为Nova模型
+	isNova, _ := c.Get("is_nova_model")
+	if isNova == true {
+		// Nova模型也支持跨区域
+		awsRegionPrefix := awsRegionPrefix(awsCli.Options().Region)
+		canCrossRegion := awsModelCanCrossRegion(awsModelId, awsRegionPrefix)
+		if canCrossRegion {
+			awsModelId = awsModelCrossRegion(awsModelId, awsRegionPrefix)
+		}
+		return handleNovaRequest(c, awsCli, info, awsModelId)
+	}
 
+	// 原有的Claude处理逻辑
 	awsRegionPrefix := awsRegionPrefix(awsCli.Options().Region)
 	canCrossRegion := awsModelCanCrossRegion(awsModelId, awsRegionPrefix)
 	if canCrossRegion {
@@ -209,3 +222,74 @@ func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rel
 	claude.HandleStreamFinalResponse(c, info, claudeInfo, RequestModeMessage)
 	return nil, claudeInfo.Usage
 }
+
+// Nova模型处理函数
+func handleNovaRequest(c *gin.Context, awsCli *bedrockruntime.Client, info *relaycommon.RelayInfo, awsModelId string) (*types.NewAPIError, *dto.Usage) {
+	novaReq_, ok := c.Get("converted_request")
+	if !ok {
+		return types.NewError(errors.New("nova request not found"), types.ErrorCodeInvalidRequest), nil
+	}
+	novaReq := novaReq_.(*NovaRequest)
+
+	// 使用InvokeModel API，但使用Nova格式的请求体
+	awsReq := &bedrockruntime.InvokeModelInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	reqBody, err := json.Marshal(novaReq)
+	if err != nil {
+		return types.NewError(errors.Wrap(err, "marshal nova request"), types.ErrorCodeBadResponseBody), nil
+	}
+	awsReq.Body = reqBody
+
+	awsResp, err := awsCli.InvokeModel(c.Request.Context(), awsReq)
+	if err != nil {
+		return types.NewError(errors.Wrap(err, "InvokeModel"), types.ErrorCodeChannelAwsClientError), nil
+	}
+
+	// 解析Nova响应
+	var novaResp struct {
+		Output struct {
+			Message struct {
+				Content []struct {
+					Text string `json:"text"`
+				} `json:"content"`
+			} `json:"message"`
+		} `json:"output"`
+		Usage struct {
+			InputTokens  int `json:"inputTokens"`
+			OutputTokens int `json:"outputTokens"`
+			TotalTokens  int `json:"totalTokens"`
+		} `json:"usage"`
+	}
+
+	if err := json.Unmarshal(awsResp.Body, &novaResp); err != nil {
+		return types.NewError(errors.Wrap(err, "unmarshal nova response"), types.ErrorCodeBadResponseBody), nil
+	}
+
+	// 构造OpenAI格式响应
+	response := dto.OpenAITextResponse{
+		Id:      helper.GetResponseID(c),
+		Object:  "chat.completion",
+		Created: common.GetTimestamp(),
+		Model:   info.UpstreamModelName,
+		Choices: []dto.OpenAITextResponseChoice{{
+			Index: 0,
+			Message: dto.Message{
+				Role:    "assistant",
+				Content: novaResp.Output.Message.Content[0].Text,
+			},
+			FinishReason: "stop",
+		}},
+		Usage: dto.Usage{
+			PromptTokens:     novaResp.Usage.InputTokens,
+			CompletionTokens: novaResp.Usage.OutputTokens,
+			TotalTokens:      novaResp.Usage.TotalTokens,
+		},
+	}
+
+	c.JSON(http.StatusOK, response)
+	return nil, &response.Usage
+}

relay/channel/vertex/adaptor.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"one-api/common"
 	"one-api/dto"
 	"one-api/relay/channel"
 	"one-api/relay/channel/claude"
@@ -80,16 +81,64 @@ func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
 	}
 }
 
-func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
-	adc := &Credentials{}
-	if err := json.Unmarshal([]byte(info.ApiKey), adc); err != nil {
-		return "", fmt.Errorf("failed to decode credentials file: %w", err)
-	}
+func (a *Adaptor) getRequestUrl(info *relaycommon.RelayInfo, modelName, suffix string) (string, error) {
 	region := GetModelRegion(info.ApiVersion, info.OriginModelName)
-	a.AccountCredentials = *adc
+	if info.ChannelOtherSettings.VertexKeyType != dto.VertexKeyTypeAPIKey {
+		adc := &Credentials{}
+		if err := common.Unmarshal([]byte(info.ApiKey), adc); err != nil {
+			return "", fmt.Errorf("failed to decode credentials file: %w", err)
+		}
+		a.AccountCredentials = *adc
+
+		if a.RequestMode == RequestModeLlama {
+			return fmt.Sprintf(
+				"https://%s-aiplatform.googleapis.com/v1beta1/projects/%s/locations/%s/endpoints/openapi/chat/completions",
+				region,
+				adc.ProjectID,
+				region,
+			), nil
+		}
+
+		if region == "global" {
+			return fmt.Sprintf(
+				"https://aiplatform.googleapis.com/v1/projects/%s/locations/global/publishers/google/models/%s:%s",
+				adc.ProjectID,
+				modelName,
+				suffix,
+			), nil
+		} else {
+			return fmt.Sprintf(
+				"https://%s-aiplatform.googleapis.com/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
+				region,
+				adc.ProjectID,
+				region,
+				modelName,
+				suffix,
+			), nil
+		}
+	} else {
+		if region == "global" {
+			return fmt.Sprintf(
+				"https://aiplatform.googleapis.com/v1/publishers/google/models/%s:%s?key=%s",
+				modelName,
+				suffix,
+				info.ApiKey,
+			), nil
+		} else {
+			return fmt.Sprintf(
+				"https://%s-aiplatform.googleapis.com/v1/publishers/google/models/%s:%s?key=%s",
+				region,
+				modelName,
+				suffix,
+				info.ApiKey,
+			), nil
+		}
+	}
+}
+
+func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
 	suffix := ""
 	if a.RequestMode == RequestModeGemini {
-
 		if model_setting.GetGeminiSettings().ThinkingAdapterEnabled {
 			// 新增逻辑：处理 -thinking-<budget> 格式
 			if strings.Contains(info.UpstreamModelName, "-thinking-") {
@@ -112,23 +161,7 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
 			suffix = "predict"
 		}
 
-		if region == "global" {
-			return fmt.Sprintf(
-				"https://aiplatform.googleapis.com/v1/projects/%s/locations/global/publishers/google/models/%s:%s",
-				adc.ProjectID,
-				info.UpstreamModelName,
-				suffix,
-			), nil
-		} else {
-			return fmt.Sprintf(
-				"https://%s-aiplatform.googleapis.com/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
-				region,
-				adc.ProjectID,
-				region,
-				info.UpstreamModelName,
-				suffix,
-			), nil
-		}
+		return a.getRequestUrl(info, info.UpstreamModelName, suffix)
 	} else if a.RequestMode == RequestModeClaude {
 		if info.IsStream {
 			suffix = "streamRawPredict?alt=sse"
@@ -139,41 +172,22 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
 		if v, ok := claudeModelMap[info.UpstreamModelName]; ok {
 			model = v
 		}
-		if region == "global" {
-			return fmt.Sprintf(
-				"https://aiplatform.googleapis.com/v1/projects/%s/locations/global/publishers/anthropic/models/%s:%s",
-				adc.ProjectID,
-				model,
-				suffix,
-			), nil
-		} else {
-			return fmt.Sprintf(
-				"https://%s-aiplatform.googleapis.com/v1/projects/%s/locations/%s/publishers/anthropic/models/%s:%s",
-				region,
-				adc.ProjectID,
-				region,
-				model,
-				suffix,
-			), nil
-		}
+		return a.getRequestUrl(info, model, suffix)
 	} else if a.RequestMode == RequestModeLlama {
-		return fmt.Sprintf(
-			"https://%s-aiplatform.googleapis.com/v1beta1/projects/%s/locations/%s/endpoints/openapi/chat/completions",
-			region,
-			adc.ProjectID,
-			region,
-		), nil
+		return a.getRequestUrl(info, "", "")
 	}
 	return "", errors.New("unsupported request mode")
 }
 
 func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error {
 	channel.SetupApiRequestHeader(info, c, req)
-	accessToken, err := getAccessToken(a, info)
-	if err != nil {
-		return err
+	if info.ChannelOtherSettings.VertexKeyType == "json" {
+		accessToken, err := getAccessToken(a, info)
+		if err != nil {
+			return err
+		}
+		req.Set("Authorization", "Bearer "+accessToken)
 	}
-	req.Set("Authorization", "Bearer "+accessToken)
 	return nil
 }
 

relay/claude_handler.go

@@ -111,7 +111,7 @@ func ClaudeHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *typ
 		httpResp = resp.(*http.Response)
 		info.IsStream = info.IsStream || strings.HasPrefix(httpResp.Header.Get("Content-Type"), "text/event-stream")
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError

relay/compatible_handler.go

@@ -158,7 +158,7 @@ func TextHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *types
 		httpResp = resp.(*http.Response)
 		info.IsStream = info.IsStream || strings.HasPrefix(httpResp.Header.Get("Content-Type"), "text/event-stream")
 		if httpResp.StatusCode != http.StatusOK {
-			newApiErr := service.RelayErrorHandler(httpResp, false)
+			newApiErr := service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newApiErr, statusCodeMappingStr)
 			return newApiErr
@@ -195,6 +195,8 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 	imageTokens := usage.PromptTokensDetails.ImageTokens
 	audioTokens := usage.PromptTokensDetails.AudioTokens
 	completionTokens := usage.CompletionTokens
+	cachedCreationTokens := usage.PromptTokensDetails.CachedCreationTokens
+
 	modelName := relayInfo.OriginModelName
 
 	tokenName := ctx.GetString("token_name")
@@ -204,6 +206,7 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 	modelRatio := relayInfo.PriceData.ModelRatio
 	groupRatio := relayInfo.PriceData.GroupRatioInfo.GroupRatio
 	modelPrice := relayInfo.PriceData.ModelPrice
+	cachedCreationRatio := relayInfo.PriceData.CacheCreationRatio
 
 	// Convert values to decimal for precise calculation
 	dPromptTokens := decimal.NewFromInt(int64(promptTokens))
@@ -211,12 +214,14 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 	dImageTokens := decimal.NewFromInt(int64(imageTokens))
 	dAudioTokens := decimal.NewFromInt(int64(audioTokens))
 	dCompletionTokens := decimal.NewFromInt(int64(completionTokens))
+	dCachedCreationTokens := decimal.NewFromInt(int64(cachedCreationTokens))
 	dCompletionRatio := decimal.NewFromFloat(completionRatio)
 	dCacheRatio := decimal.NewFromFloat(cacheRatio)
 	dImageRatio := decimal.NewFromFloat(imageRatio)
 	dModelRatio := decimal.NewFromFloat(modelRatio)
 	dGroupRatio := decimal.NewFromFloat(groupRatio)
 	dModelPrice := decimal.NewFromFloat(modelPrice)
+	dCachedCreationRatio := decimal.NewFromFloat(cachedCreationRatio)
 	dQuotaPerUnit := decimal.NewFromFloat(common.QuotaPerUnit)
 
 	ratio := dModelRatio.Mul(dGroupRatio)
@@ -284,6 +289,11 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 			baseTokens = baseTokens.Sub(dCacheTokens)
 			cachedTokensWithRatio = dCacheTokens.Mul(dCacheRatio)
 		}
+		var dCachedCreationTokensWithRatio decimal.Decimal
+		if !dCachedCreationTokens.IsZero() {
+			baseTokens = baseTokens.Sub(dCachedCreationTokens)
+			dCachedCreationTokensWithRatio = dCachedCreationTokens.Mul(dCachedCreationRatio)
+		}
 
 		// 减去 image tokens
 		var imageTokensWithRatio decimal.Decimal
@@ -302,7 +312,9 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 				extraContent += fmt.Sprintf("Audio Input 花费 %s", audioInputQuota.String())
 			}
 		}
-		promptQuota := baseTokens.Add(cachedTokensWithRatio).Add(imageTokensWithRatio)
+		promptQuota := baseTokens.Add(cachedTokensWithRatio).
+			Add(imageTokensWithRatio).
+			Add(dCachedCreationTokensWithRatio)
 
 		completionQuota := dCompletionTokens.Mul(dCompletionRatio)
 
@@ -395,6 +407,10 @@ func postConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage
 		other["image_ratio"] = imageRatio
 		other["image_output"] = imageTokens
 	}
+	if cachedCreationTokens != 0 {
+		other["cache_creation_tokens"] = cachedCreationTokens
+		other["cache_creation_ratio"] = cachedCreationRatio
+	}
 	if !dWebSearchQuota.IsZero() {
 		if relayInfo.ResponsesUsageInfo != nil {
 			if webSearchTool, exists := relayInfo.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolWebSearchPreview]; exists {

relay/embedding_handler.go

@@ -58,7 +58,7 @@ func EmbeddingHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError

relay/gemini_handler.go

@@ -152,7 +152,7 @@ func GeminiHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *typ
 		httpResp = resp.(*http.Response)
 		info.IsStream = info.IsStream || strings.HasPrefix(httpResp.Header.Get("Content-Type"), "text/event-stream")
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError
@@ -249,7 +249,7 @@ func GeminiEmbeddingHandler(c *gin.Context, info *relaycommon.RelayInfo) (newAPI
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError
 		}

relay/image_handler.go

@@ -91,7 +91,7 @@ func ImageHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *type
 		httpResp = resp.(*http.Response)
 		info.IsStream = info.IsStream || strings.HasPrefix(httpResp.Header.Get("Content-Type"), "text/event-stream")
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError
@@ -120,7 +120,7 @@ func ImageHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *type
 	var logContent string
 
 	if len(request.Size) > 0 {
-		logContent = fmt.Sprintf("大小 %s, 品质 %s", request.Size, quality)
+		logContent = fmt.Sprintf("大小 %s, 品质 %s, 张数 %d", request.Size, quality, request.N)
 	}
 
 	postConsumeQuota(c, info, usage.(*dto.Usage), logContent)

relay/rerank_handler.go

@@ -81,7 +81,7 @@ func RerankHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *typ
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError

relay/responses_handler.go

@@ -41,7 +41,7 @@ func ResponsesHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *
 	}
 	adaptor.Init(info)
 	var requestBody io.Reader
-	if model_setting.GetGlobalSettings().PassThroughRequestEnabled {
+	if model_setting.GetGlobalSettings().PassThroughRequestEnabled || info.ChannelSetting.PassThroughBodyEnabled {
 		body, err := common.GetRequestBody(c)
 		if err != nil {
 			return types.NewError(err, types.ErrorCodeReadRequestBodyFailed, types.ErrOptionWithSkipRetry())
@@ -82,7 +82,7 @@ func ResponsesHelper(c *gin.Context, info *relaycommon.RelayInfo) (newAPIError *
 		httpResp = resp.(*http.Response)
 
 		if httpResp.StatusCode != http.StatusOK {
-			newAPIError = service.RelayErrorHandler(httpResp, false)
+			newAPIError = service.RelayErrorHandler(c.Request.Context(), httpResp, false)
 			// reset status code 重置状态码
 			service.ResetStatusCode(newAPIError, statusCodeMappingStr)
 			return newAPIError

service/error.go

@@ -1,12 +1,14 @@
 package service
 
 import (
+	"context"
 	"errors"
 	"fmt"
 	"io"
 	"net/http"
 	"one-api/common"
 	"one-api/dto"
+	"one-api/logger"
 	"one-api/types"
 	"strconv"
 	"strings"
@@ -78,7 +80,7 @@ func ClaudeErrorWrapperLocal(err error, code string, statusCode int) *dto.Claude
 	return claudeErr
 }
 
-func RelayErrorHandler(resp *http.Response, showBodyWhenFail bool) (newApiErr *types.NewAPIError) {
+func RelayErrorHandler(ctx context.Context, resp *http.Response, showBodyWhenFail bool) (newApiErr *types.NewAPIError) {
 	newApiErr = types.InitOpenAIError(types.ErrorCodeBadResponseStatusCode, resp.StatusCode)
 
 	responseBody, err := io.ReadAll(resp.Body)
@@ -94,7 +96,7 @@ func RelayErrorHandler(resp *http.Response, showBodyWhenFail bool) (newApiErr *t
 			newApiErr.Err = fmt.Errorf("bad response status code %d, body: %s", resp.StatusCode, string(responseBody))
 		} else {
 			if common.DebugEnabled {
-				println(fmt.Sprintf("bad response status code %d, body: %s", resp.StatusCode, string(responseBody)))
+				logger.LogInfo(ctx, fmt.Sprintf("bad response status code %d, body: %s", resp.StatusCode, string(responseBody)))
 			}
 			newApiErr.Err = fmt.Errorf("bad response status code %d", resp.StatusCode)
 		}

service/pre_consume_quota.go

@@ -13,13 +13,13 @@ import (
 	"github.com/gin-gonic/gin"
 )
 
-func ReturnPreConsumedQuota(c *gin.Context, relayInfo *relaycommon.RelayInfo, preConsumedQuota int) {
-	if preConsumedQuota != 0 {
-		logger.LogInfo(c, fmt.Sprintf("用户 %d 请求失败, 返还预扣费额度 %s", relayInfo.UserId, logger.FormatQuota(preConsumedQuota)))
+func ReturnPreConsumedQuota(c *gin.Context, relayInfo *relaycommon.RelayInfo) {
+	if relayInfo.FinalPreConsumedQuota != 0 {
+		logger.LogInfo(c, fmt.Sprintf("用户 %d 请求失败, 返还预扣费额度 %s", relayInfo.UserId, logger.FormatQuota(relayInfo.FinalPreConsumedQuota)))
 		gopool.Go(func() {
 			relayInfoCopy := *relayInfo
 
-			err := PostConsumeQuota(&relayInfoCopy, -preConsumedQuota, 0, false)
+			err := PostConsumeQuota(&relayInfoCopy, -relayInfo.FinalPreConsumedQuota, 0, false)
 			if err != nil {
 				common.SysLog("error return pre-consumed quota: " + err.Error())
 			}
@@ -29,16 +29,16 @@ func ReturnPreConsumedQuota(c *gin.Context, relayInfo *relaycommon.RelayInfo, pr
 
 // PreConsumeQuota checks if the user has enough quota to pre-consume.
 // It returns the pre-consumed quota if successful, or an error if not.
-func PreConsumeQuota(c *gin.Context, preConsumedQuota int, relayInfo *relaycommon.RelayInfo) (int, *types.NewAPIError) {
+func PreConsumeQuota(c *gin.Context, preConsumedQuota int, relayInfo *relaycommon.RelayInfo) *types.NewAPIError {
 	userQuota, err := model.GetUserQuota(relayInfo.UserId, false)
 	if err != nil {
-		return 0, types.NewError(err, types.ErrorCodeQueryDataError, types.ErrOptionWithSkipRetry())
+		return types.NewError(err, types.ErrorCodeQueryDataError, types.ErrOptionWithSkipRetry())
 	}
 	if userQuota <= 0 {
-		return 0, types.NewErrorWithStatusCode(fmt.Errorf("用户额度不足, 剩余额度: %s", logger.FormatQuota(userQuota)), types.ErrorCodeInsufficientUserQuota, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
+		return types.NewErrorWithStatusCode(fmt.Errorf("用户额度不足, 剩余额度: %s", logger.FormatQuota(userQuota)), types.ErrorCodeInsufficientUserQuota, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
 	}
 	if userQuota-preConsumedQuota < 0 {
-		return 0, types.NewErrorWithStatusCode(fmt.Errorf("预扣费额度失败, 用户剩余额度: %s, 需要预扣费额度: %s", logger.FormatQuota(userQuota), logger.FormatQuota(preConsumedQuota)), types.ErrorCodeInsufficientUserQuota, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
+		return types.NewErrorWithStatusCode(fmt.Errorf("预扣费额度失败, 用户剩余额度: %s, 需要预扣费额度: %s", logger.FormatQuota(userQuota), logger.FormatQuota(preConsumedQuota)), types.ErrorCodeInsufficientUserQuota, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
 	}
 
 	trustQuota := common.GetTrustQuota()
@@ -65,14 +65,14 @@ func PreConsumeQuota(c *gin.Context, preConsumedQuota int, relayInfo *relaycommo
 	if preConsumedQuota > 0 {
 		err := PreConsumeTokenQuota(relayInfo, preConsumedQuota)
 		if err != nil {
-			return 0, types.NewErrorWithStatusCode(err, types.ErrorCodePreConsumeTokenQuotaFailed, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
+			return types.NewErrorWithStatusCode(err, types.ErrorCodePreConsumeTokenQuotaFailed, http.StatusForbidden, types.ErrOptionWithSkipRetry(), types.ErrOptionWithNoRecordErrorLog())
 		}
 		err = model.DecreaseUserQuota(relayInfo.UserId, preConsumedQuota)
 		if err != nil {
-			return 0, types.NewError(err, types.ErrorCodeUpdateDataError, types.ErrOptionWithSkipRetry())
+			return types.NewError(err, types.ErrorCodeUpdateDataError, types.ErrOptionWithSkipRetry())
 		}
 		logger.LogInfo(c, fmt.Sprintf("用户 %d 预扣费 %s, 预扣费后剩余额度: %s", relayInfo.UserId, logger.FormatQuota(preConsumedQuota), logger.FormatQuota(userQuota-preConsumedQuota)))
 	}
 	relayInfo.FinalPreConsumedQuota = preConsumedQuota
-	return preConsumedQuota, nil
+	return nil
 }

types/error.go

@@ -185,6 +185,14 @@ func (e *NewAPIError) ToClaudeError() ClaudeError {
 type NewAPIErrorOptions func(*NewAPIError)
 
 func NewError(err error, errorCode ErrorCode, ops ...NewAPIErrorOptions) *NewAPIError {
+	var newErr *NewAPIError
+	// 保留深层传递的 new err
+	if errors.As(err, &newErr) {
+		for _, op := range ops {
+			op(newErr)
+		}
+		return newErr
+	}
 	e := &NewAPIError{
 		Err:        err,
 		RelayError: nil,
@@ -199,8 +207,21 @@ func NewError(err error, errorCode ErrorCode, ops ...NewAPIErrorOptions) *NewAPI
 }
 
 func NewOpenAIError(err error, errorCode ErrorCode, statusCode int, ops ...NewAPIErrorOptions) *NewAPIError {
-	if errorCode == ErrorCodeDoRequestFailed {
-		err = errors.New("upstream error: do request failed")
+	var newErr *NewAPIError
+	// 保留深层传递的 new err
+	if errors.As(err, &newErr) {
+		if newErr.RelayError == nil {
+			openaiError := OpenAIError{
+				Message: newErr.Error(),
+				Type:    string(errorCode),
+				Code:    errorCode,
+			}
+			newErr.RelayError = openaiError
+		}
+		for _, op := range ops {
+			op(newErr)
+		}
+		return newErr
 	}
 	openaiError := OpenAIError{
 		Message: err.Error(),
@@ -305,6 +326,15 @@ func ErrOptionWithNoRecordErrorLog() NewAPIErrorOptions {
 	}
 }
 
+func ErrOptionWithHideErrMsg(replaceStr string) NewAPIErrorOptions {
+	return func(e *NewAPIError) {
+		if common.DebugEnabled {
+			fmt.Printf("ErrOptionWithHideErrMsg: %s, origin error: %s", replaceStr, e.Err)
+		}
+		e.Err = errors.New(replaceStr)
+	}
+}
+
 func IsRecordErrorLog(e *NewAPIError) bool {
 	if e == nil {
 		return false

web/src/components/table/channels/modals/EditChannelModal.jsx

@@ -142,6 +142,8 @@ const EditChannelModal = (props) => {
     system_prompt: '',
     system_prompt_override: false,
     settings: '',
+    // 仅 Vertex: 密钥格式（存入 settings.vertex_key_type）
+    vertex_key_type: 'json',
   };
   const [batch, setBatch] = useState(false);
   const [multiToSingle, setMultiToSingle] = useState(false);
@@ -409,11 +411,17 @@ const EditChannelModal = (props) => {
           const parsedSettings = JSON.parse(data.settings);
           data.azure_responses_version =
             parsedSettings.azure_responses_version || '';
+          // 读取 Vertex 密钥格式
+          data.vertex_key_type = parsedSettings.vertex_key_type || 'json';
         } catch (error) {
           console.error('解析其他设置失败:', error);
           data.azure_responses_version = '';
           data.region = '';
+          data.vertex_key_type = 'json';
         }
+      } else {
+        // 兼容历史数据：老渠道没有 settings 时，默认按 json 展示
+        data.vertex_key_type = 'json';
       }
 
       setInputs(data);
@@ -745,59 +753,56 @@ const EditChannelModal = (props) => {
     let localInputs = { ...formValues };
 
     if (localInputs.type === 41) {
-      if (useManualInput) {
-        // 手动输入模式
-        if (localInputs.key && localInputs.key.trim() !== '') {
-          try {
-            // 验证 JSON 格式
-            const parsedKey = JSON.parse(localInputs.key);
-            // 确保是有效的密钥格式
-            localInputs.key = JSON.stringify(parsedKey);
-          } catch (err) {
-            showError(t('密钥格式无效，请输入有效的 JSON 格式密钥'));
-            return;
-          }
-        } else if (!isEdit) {
+      const keyType = localInputs.vertex_key_type || 'json';
+      if (keyType === 'api_key') {
+        // 直接作为普通字符串密钥处理
+        if (!isEdit && (!localInputs.key || localInputs.key.trim() === '')) {
           showInfo(t('请输入密钥！'));
           return;
         }
       } else {
-        // 文件上传模式
-        let keys = vertexKeys;
-
-        // 若当前未选择文件，尝试从已上传文件列表解析（异步读取）
-        if (keys.length === 0 && vertexFileList.length > 0) {
-          try {
-            const parsed = await Promise.all(
-              vertexFileList.map(async (item) => {
-                const fileObj = item.fileInstance;
-                if (!fileObj) return null;
-                const txt = await fileObj.text();
-                return JSON.parse(txt);
-              }),
-            );
-            keys = parsed.filter(Boolean);
-          } catch (err) {
-            showError(t('解析密钥文件失败: {{msg}}', { msg: err.message }));
+        // JSON 服务账号密钥
+        if (useManualInput) {
+          if (localInputs.key && localInputs.key.trim() !== '') {
+            try {
+              const parsedKey = JSON.parse(localInputs.key);
+              localInputs.key = JSON.stringify(parsedKey);
+            } catch (err) {
+              showError(t('密钥格式无效，请输入有效的 JSON 格式密钥'));
+              return;
+            }
+          } else if (!isEdit) {
+            showInfo(t('请输入密钥！'));
             return;
           }
-        }
-
-        // 创建模式必须上传密钥；编辑模式可选
-        if (keys.length === 0) {
-          if (!isEdit) {
-            showInfo(t('请上传密钥文件！'));
-            return;
-          } else {
-            // 编辑模式且未上传新密钥，不修改 key
-            delete localInputs.key;
-          }
         } else {
-          // 有新密钥，则覆盖
-          if (batch) {
-            localInputs.key = JSON.stringify(keys);
+          // 文件上传模式
+          let keys = vertexKeys;
+          if (keys.length === 0 && vertexFileList.length > 0) {
+            try {
+              const parsed = await Promise.all(
+                vertexFileList.map(async (item) => {
+                  const fileObj = item.fileInstance;
+                  if (!fileObj) return null;
+                  const txt = await fileObj.text();
+                  return JSON.parse(txt);
+                }),
+              );
+              keys = parsed.filter(Boolean);
+            } catch (err) {
+              showError(t('解析密钥文件失败: {{msg}}', { msg: err.message }));
+              return;
+            }
+          }
+          if (keys.length === 0) {
+            if (!isEdit) {
+              showInfo(t('请上传密钥文件！'));
+              return;
+            } else {
+              delete localInputs.key;
+            }
           } else {
-            localInputs.key = JSON.stringify(keys[0]);
+            localInputs.key = batch ? JSON.stringify(keys) : JSON.stringify(keys[0]);
           }
         }
       }
@@ -853,6 +858,8 @@ const EditChannelModal = (props) => {
     delete localInputs.pass_through_body_enabled;
     delete localInputs.system_prompt;
     delete localInputs.system_prompt_override;
+    // 顶层的 vertex_key_type 不应发送给后端
+    delete localInputs.vertex_key_type;
 
     let res;
     localInputs.auto_ban = localInputs.auto_ban ? 1 : 0;
@@ -1178,8 +1185,40 @@ const EditChannelModal = (props) => {
                     autoComplete='new-password'
                   />
 
+                  {inputs.type === 41 && (
+                    <Form.Select
+                      field='vertex_key_type'
+                      label={t('密钥格式')}
+                      placeholder={t('请选择密钥格式')}
+                      optionList={[
+                        { label: 'JSON', value: 'json' },
+                        { label: 'API Key', value: 'api_key' },
+                      ]}
+                      style={{ width: '100%' }}
+                      value={inputs.vertex_key_type || 'json'}
+                      onChange={(value) => {
+                        // 更新设置中的 vertex_key_type
+                        handleChannelOtherSettingsChange('vertex_key_type', value);
+                        // 切换为 api_key 时，关闭批量与手动/文件切换，并清理已选文件
+                        if (value === 'api_key') {
+                          setBatch(false);
+                          setUseManualInput(false);
+                          setVertexKeys([]);
+                          setVertexFileList([]);
+                          if (formApiRef.current) {
+                            formApiRef.current.setValue('vertex_files', []);
+                          }
+                        }
+                      }}
+                      extraText={
+                        inputs.vertex_key_type === 'api_key'
+                          ? t('API Key 模式下不支持批量创建')
+                          : t('JSON 模式支持手动输入或上传服务账号 JSON')
+                      }
+                    />
+                  )}
                   {batch ? (
-                    inputs.type === 41 ? (
+                    inputs.type === 41 && (inputs.vertex_key_type || 'json') === 'json' ? (
                       <Form.Upload
                         field='vertex_files'
                         label={t('密钥文件 (.json)')}
@@ -1243,7 +1282,7 @@ const EditChannelModal = (props) => {
                     )
                   ) : (
                     <>
-                      {inputs.type === 41 ? (
+                      {inputs.type === 41 && (inputs.vertex_key_type || 'json') === 'json' ? (
                         <>
                           {!batch && (
                             <div className='flex items-center justify-between mb-3'>

web/src/pages/Setting/Operation/SettingsGeneral.jsx

@@ -194,7 +194,7 @@ export default function GeneralSettings(props) {
                 />
               </Col>
             </Row>
-            <Row>
+            <Row gutter={16}>
               <Col xs={24} sm={12} md={8} lg={8} xl={8}>
                 <Form.Switch
                   field={'DemoSiteEnabled'}

web/src/pages/Setting/Operation/SettingsHeaderNavModules.jsx

@@ -304,7 +304,7 @@ export default function SettingsHeaderNavModules(props) {
                               headerNavModules.pricing?.requireAuth || false
                             }
                             onChange={handlePricingAuthChange}
-                            size='small'
+                            size='default'
                           />
                         </div>
                       </div>