adapt for llama server

tradingagents/default_config.py

@@ -11,7 +11,7 @@ DEFAULT_CONFIG = {
     "llm_provider": "openai",
     "deep_think_llm": "qwen-35b",
     "quick_think_llm": "qwen-35b",
-    "backend_url": "https://ai.ghosthouse.uk/v1/chat/completions",
+    "backend_url": "https://ai.ghosthouse.uk/v1",
     # Provider-specific thinking configuration
     "google_thinking_level": None,      # "high", "minimal", etc.
     "openai_reasoning_effort": None,    # "medium", "high", "low"

tradingagents/llm_clients/openai_client.py

@@ -76,7 +76,7 @@ class OpenAIClient(BaseLLMClient):
 
         # Native OpenAI: use Responses API for consistent behavior across
         # all model families. Third-party providers use Chat Completions.
-        if self.provider == "openai":
+        if self.provider == "openai" and not self.base_url:
             llm_kwargs["use_responses_api"] = True
 
         return NormalizedChatOpenAI(**llm_kwargs)