fix: update OLLAMA_API_URL to correct endpoint and add OLLAMA_MODEL environment variable

.env.example

@@ -12,3 +12,7 @@ OIDC_VERIFY_ISS=true
 
 # Frontend redirect URI
 VITE_OIDC_REDIRECT_URI=http://localhost:5173
+
+# Ollama Cloud model name (optional)
+# Set to a valid Ollama model for your account, e.g. "llama3" or "gpt-4o-mini".
+OLLAMA_MODEL=llama3

backend/app/main.py

@@ -11,7 +11,8 @@ from app.auth import get_discovery, settings, validate_token
 from app.schemas import AuthCallbackRequest, AuthConfigResponse, ChatRequest
 
 OLLAMA_API_KEY = os.getenv("OLLAMA_API_KEY")
-OLLAMA_API_URL = os.getenv("OLLAMA_API_URL", "https://ollama.com/v1/chat/completions")
+OLLAMA_API_URL = os.getenv("OLLAMA_API_URL", "https://api.ollama.com/v1/chat/completions")
+OLLAMA_MODEL = os.getenv("OLLAMA_MODEL")
 if not OLLAMA_API_KEY:
     raise RuntimeError("Environment variable OLLAMA_API_KEY is required")
 
@@ -33,11 +34,13 @@ app.add_middleware(
 )
 
 async def event_stream(request: ChatRequest) -> AsyncGenerator[str, None]:
+    model = OLLAMA_MODEL or request.model
     payload = {
-        "model": request.model,
+        "model": model,
         "messages": [message.dict() for message in request.messages],
         "stream": True,
     }
+    logger.debug("Sending Ollama request with model=%s", model)
     headers = {
         "Authorization": f"Bearer {OLLAMA_API_KEY}",
         "Content-Type": "application/json",