fix(chat-api): support Modal inference URL in inference client

The inference client now auto-detects if the URL already ends with
/generate (Modal's endpoint URL pattern) and skips appending the path.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

services/chat-api/src/services/inference_client.py

@@ -84,9 +84,15 @@ class InferenceClient:
         }
 
         client = self._get_client()
+        # If the base_url already ends with a path (e.g. Modal endpoint URL),
+        # use it directly. Otherwise append /generate for the local service.
+        url = self.base_url
+        if not url.endswith("/generate"):
+            url = f"{url}/generate"
+
         async with client.stream(
             "POST",
-            f"{self.base_url}/generate",
+            url,
             headers=self.headers,
             json=payload,
         ) as response: