diff --git a/README.md b/README.md
index 323562e..f6f3f53 100644
--- a/README.md
+++ b/README.md
@@ -75,6 +75,12 @@ The server currently supports the following endpoints:
 > **Note**: `/v1/completions` is not supported because Gemini models do not support completion functionality, and Google's PaLM model (which does) is likely to be deprecated.  
 
 
+## Roadmap
+
+- [x] v1
+- [ ] 支持配置多个 api key ，并且可以轮询、分配权限、重试等调度操作
+- [ ] 支持配置和切换代理
+
 ## Building the Project  
 
 ```
diff --git a/package.json b/package.json
index a58b635..c5aa53e 100644
--- a/package.json
+++ b/package.json
@@ -12,6 +12,7 @@
     "@fastify/swagger": "^8.15.0",
     "@fastify/swagger-ui": "^2.1.0",
     "@google/generative-ai": "^0.21.0",
+    "@types/async-retry": "^1.4.9",
     "async-retry": "^1.3.3",
     "commander": "^12.1.0",
     "dotenv": "^16.4.5",
diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index f5d22b4..955b03a 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -17,6 +17,9 @@ dependencies:
   '@google/generative-ai':
     specifier: ^0.21.0
     version: 0.21.0
+  '@types/async-retry':
+    specifier: ^1.4.9
+    version: 1.4.9
   async-retry:
     specifier: ^1.3.3
     version: 1.3.3
@@ -1975,6 +1978,12 @@ packages:
     resolution: {integrity: sha512-C5Mc6rdnsaJDjO3UpGW/CQTHtCKaYlScZTly4JIu97Jxo/odCiH0ITnDXSJPTOrEKk/ycSZ0AOgTmkDtkOsvIA==}
     dev: false
 
+  /@types/async-retry@1.4.9:
+    resolution: {integrity: sha512-s1ciZQJzRh3708X/m3vPExr5KJlzlZJvXsKpbtE2luqNcbROr64qU+3KpJsYHqWMeaxI839OvXf9PrUSw1Xtyg==}
+    dependencies:
+      '@types/retry': 0.12.5
+    dev: false
+
   /@types/cacheable-request@6.0.3:
     resolution: {integrity: sha512-IQ3EbTzGxIigb1I3qPZc1rWJnH0BmSKv5QYTalEwweFvyBDLSAe24zP0le/hyi7ecGfZVlIVAg4BZqb8WBwKqw==}
     dependencies:
@@ -2029,6 +2038,10 @@ packages:
       '@types/node': 22.10.1
     dev: true
 
+  /@types/retry@0.12.5:
+    resolution: {integrity: sha512-3xSjTp3v03X/lSQLkczaN9UIEwJMoMCA1+Nb5HfbJEQWogdeQIyVtTvxPXDQjZ5zws8rFQfVfRdz03ARihPJgw==}
+    dev: false
+
   /@types/triple-beam@1.3.5:
     resolution: {integrity: sha512-6WaYesThRMCl19iryMYP7/x2OVgCtbIVflDGFpWnb9irXI3UjYE4AzmYuiUKY1AJstGijoY+MgUszMgRxIYTYw==}
     dev: true
diff --git a/src/v1/chat/completions.ts b/src/v1/chat/completions.ts
index de32b87..4280ef7 100644
--- a/src/v1/chat/completions.ts
+++ b/src/v1/chat/completions.ts
@@ -192,6 +192,7 @@ class ChatCompletionsHandler {
         const is_hit_err = retry_error.some((x) => err instanceof x);
         if (is_hit_err === false) {
           bail(err);
+          if (err) err.bail = true;
         }
         throw err; // 触发重试
       }
@@ -519,7 +520,12 @@ const chatCompletionsRoute: FastifyPluginAsync = async (app) => {
     url: "/v1/chat/completions",
     schema: { body: CHAT_COMPLETION_SCHEMA },
     handler: async (req, reply) => {
-      const handler = new ChatCompletionsHandler(app, req.body, req, reply);
+      const handler = new ChatCompletionsHandler(
+        app,
+        req.body as any,
+        req,
+        reply
+      );
       await handler.processRequest();
     },
   });