diff --git a/relay/adaptor/doubao/main.go b/relay/adaptor/doubao/main.go
index dd43d06c53..342b966a9e 100644
--- a/relay/adaptor/doubao/main.go
+++ b/relay/adaptor/doubao/main.go
@@ -7,9 +7,13 @@ import (
 )
 
 func GetRequestURL(meta *meta.Meta) (string, error) {
+	var context = ""
+	if meta.Cache {
+		context = "context/"
+	}
 	switch meta.Mode {
 	case relaymode.ChatCompletions:
-		return fmt.Sprintf("%s/api/v3/chat/completions", meta.BaseURL), nil
+		return fmt.Sprintf("%s/api/v3/%schat/completions", meta.BaseURL, context), nil
 	case relaymode.Embeddings:
 		return fmt.Sprintf("%s/api/v3/embeddings", meta.BaseURL), nil
 	default:
diff --git a/relay/controller/text.go b/relay/controller/text.go
index 9a47c58bc2..9e6456351a 100644
--- a/relay/controller/text.go
+++ b/relay/controller/text.go
@@ -31,6 +31,7 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 		return openai.ErrorWrapper(err, "invalid_text_request", http.StatusBadRequest)
 	}
 	meta.IsStream = textRequest.Stream
+	meta.Cache = len(textRequest.ContextId) > 0
 
 	// map model name
 	meta.OriginModelName = textRequest.Model
diff --git a/relay/meta/relay_meta.go b/relay/meta/relay_meta.go
index bcbe10453a..b33a2f99e4 100644
--- a/relay/meta/relay_meta.go
+++ b/relay/meta/relay_meta.go
@@ -31,6 +31,7 @@ type Meta struct {
 	RequestURLPath  string
 	PromptTokens    int // only for DoResponse
 	SystemPrompt    string
+	Cache           bool
 }
 
 func GetByContext(c *gin.Context) *Meta {
diff --git a/relay/model/general.go b/relay/model/general.go
index 288c07ffb5..f8fbe82d0a 100644
--- a/relay/model/general.go
+++ b/relay/model/general.go
@@ -65,6 +65,8 @@ type GeneralOpenAIRequest struct {
 	// Others
 	Instruction string `json:"instruction,omitempty"`
 	NumCtx      int    `json:"num_ctx,omitempty"`
+	// SessionCache or PrefixCache context id
+	ContextId string `json:"context_id,omitempty"`
 }
 
 func (r GeneralOpenAIRequest) ParseInput() []string {