POST: /api/chat doesn't reply.

[luckiman]

I forked project in local and set all environment variables. but it occurs error while sending message. What's wrong?

[generativelabs]

I ran into the same thing. The app/(chat)/api/chat/route.ts is eating the meaningful error, if you add:

  } catch (error) {
    console.error("[DEBUG] API error:", error);
    return new Response('An error occurred while processing your request!', {
      status: 404,
    });

you'll see what the problem is. For me, it was I didn't have XAI_API_KEY defined. The latest versions use xAI as you can see in providers.ts

[luckiman]

I got a xAI api key and set it in local.env. how to validate that api key?

[generativelabs]

how to validate that api key?

Set a breakpoint on that exception or add debug message shown above to validate / rule out an invalid key.

[cssagogo]

Running into the same issue.

[luckiman]

I checked the point occuring issue. It caused at generateTitleFromUserMessage in /api/chat/route.ts.
how to solve the "server is not available"?
Can't I use free xAi free api key?

[arthtyagi]

You need credits on xAI API key.

[luckiman]

How to get credits on xAI API key? I have already got a xAI_API_KEY in https://console.x.ai/.
But it occurs above error. why and how to fix it?

[zzkkui]

I am using deepseek and my deepseek has a credit limit. When I start it, the command line reports an error and chat returns 404.

before chat

The key used here is usable because I can chat normally in other places

[luckiman]

You mean that I should use google or deepseek api key, not a xAi key?

[GitTom]

Most of the other providers are fairly generous with their free tiers so you don't have to get into paying.

Here's a diff of changes I made to providers.ts and the package.json to switch this code template back to OpenAI.

(If you aren't familiar with this diff format then just give this to your LLM and it will help you understand it, validate it, and apply it.)

diff --git a/lib/ai/providers.ts b/lib/ai/providers.ts
index 810c329..c10b211 100644
--- a/lib/ai/providers.ts
+++ b/lib/ai/providers.ts
@@ -3,7 +3,7 @@ import {
   extractReasoningMiddleware,
   wrapLanguageModel,
 } from 'ai';
-import { xai } from '@ai-sdk/xai';
+import { openai } from '@ai-sdk/openai';
 import { isTestEnvironment } from '../constants';
 import {
   artifactModel,
@@ -23,15 +23,16 @@ export const myProvider = isTestEnvironment
     })
   : customProvider({
       languageModels: {
-        'chat-model': xai('grok-2-1212'),
+        'chat-model': openai('gpt-4o-mini'),
         'chat-model-reasoning': wrapLanguageModel({
-          model: xai('grok-3-mini-beta'),
+          model: openai('o3-mini'),
           middleware: extractReasoningMiddleware({ tagName: 'think' }),
         }),
-        'title-model': xai('grok-2-1212'),
-        'artifact-model': xai('grok-2-1212'),
+        'title-model': openai('o3-mini'),
+        'artifact-model': openai('o3-mini'),
       },
       imageModels: {
-        'small-model': xai.image('grok-2-image'),
+       'small-model': openai.image('dall-e-2'),
+       'large-model': openai.image('dall-e-3'),
       },
     });

diff --git a/package.json b/package.json
index a48d3e3..6236ef9 100644
--- a/package.json
+++ b/package.json
@@ -20,7 +20,7 @@
   },
   "dependencies": {
     "@ai-sdk/react": "^1.2.8",
-    "@ai-sdk/xai": "^1.2.10",
+    "@ai-sdk/openai": "^1.3.10",
     "@codemirror/lang-javascript": "^6.2.2",
     "@codemirror/lang-python": "^6.1.6",
     "@codemirror/state": "^6.5.0",
@@ -104,4 +104,4 @@
     "typescript": "^5.6.3"
   },
   "packageManager": "pnpm@9.12.3"
-}
+}

[luckiman]

I updated code with openAI model and added a OPENAI_API_KEY. But it occurs a similar error.

[blexo]

Also... they need to update their documentation...

[yuxizhe]

you can add console.error at app/(chat)/api/chat/route.ts line 223 & 239

https://github.com/vercel/ai-chatbot/blob/main/app/(chat)/api/chat/route.ts#L223