add HF_INFERENCE_PROVIDER

app.py

@@ -44,9 +44,11 @@ _monkeypatch_gradio_save_history()
 
 def get_agent_model(model_name: str, oauth_token: gr.OAuthToken = None):
     token = os.getenv("HF_TOKEN", None) or oauth_token.token if oauth_token else None
+    provider = os.getenv("HF_INFERENCE_PROVIDER", "together")
     model = InferenceClientModel(
         token=token,
         model_id=model_name,
+        provider=provider, 
     )
     return model
 

src/upgrade_advisor/config.py

@@ -5,8 +5,8 @@ from dotenv import load_dotenv
 load_dotenv()
 
 HF_INFERENCE_PROVIDER= os.getenv("HF_INFERENCE_PROVIDER", "together")
-AGENT_MODEL = os.getenv("AGENT_MODEL", f"Qwen/Qwen3-Next-80B-A3B-Thinking:{HF_INFERENCE_PROVIDER}")
-CHAT_MODEL = os.getenv("CHAT_MODEL", f"Qwen/Qwen3-Next-80B-A3B-Thinking:{HF_INFERENCE_PROVIDER}")
+AGENT_MODEL = os.getenv("AGENT_MODEL", f"Qwen/Qwen3-Next-80B-A3B-Thinking")
+CHAT_MODEL = os.getenv("CHAT_MODEL", f"Qwen/Qwen3-Next-80B-A3B-Thinking")
 
 # GitHub MCP configuration
 GITHUB_PAT = os.getenv("GITHUB_PAT", None)