Merge 8b1289f5fc into 589b351f2a

.env.example

@@ -1,6 +1,52 @@
-# LLM Providers (set the one you use)
+# =============================================================================
+# TradingAgents Environment Configuration
+# =============================================================================
+# Copy this file to .env and fill in your API keys and configuration
+
+# =============================================================================
+# LLM Provider API Keys
+# =============================================================================
+# Set the API key for the LLM provider you want to use
+# The CLI will prompt you to select a provider when running analysis
+
+# OpenAI (GPT-4, GPT-5 models)
 OPENAI_API_KEY=
+
+# Google (Gemini models)
 GOOGLE_API_KEY=
+
+# Anthropic (Claude models)
 ANTHROPIC_API_KEY=
+
+# xAI (Grok models)
 XAI_API_KEY=
+
+# OpenRouter (Access to multiple models)
 OPENROUTER_API_KEY=
+
+# =============================================================================
+# LM Studio Configuration (Local Models)
+# =============================================================================
+# LM Studio allows you to run models locally on your machine
+# The CLI will automatically detect available models when you select LM Studio
+
+# Base URL for LM Studio server (optional, defaults to http://localhost:1234/v1)
+# Change this if you're running LM Studio on a different port or remote server
+LMSTUDIO_BASE_URL=http://localhost:1234/v1
+
+# API Key for LM Studio (optional, only needed if you secured your instance)
+# Most local LM Studio instances don't require an API key
+LMSTUDIO_API_KEY=
+
+# =============================================================================
+# Data Provider API Keys
+# =============================================================================
+# Alpha Vantage - Stock market data, news, and fundamentals
+# Get your free API key at: https://www.alphavantage.co/support/#api-key
+ALPHA_VANTAGE_API_KEY=
+
+# =============================================================================
+# Advanced Configuration (Optional)
+# =============================================================================
+# Custom results directory (defaults to ./results)
+# TRADINGAGENTS_RESULTS_DIR=./results

.gitignore

@@ -217,3 +217,7 @@ __marimo__/
 
 # Cache
 **/data_cache/
+
+# Results
+results/*
+reports/*

cli/main.py

@@ -585,6 +585,7 @@ def get_user_selections():
             )
         )
         anthropic_effort = ask_anthropic_effort()
+    # LM Studio, Ollama, xAI, and Openrouter don't need additional configuration
 
     return {
         "ticker": selected_ticker,

cli/utils.py

@@ -1,3 +1,5 @@
+import os
+import requests
 import questionary
 from typing import List, Optional, Tuple, Dict
 
@@ -133,6 +135,31 @@ def select_research_depth() -> int:
     return choice
 
 
+def fetch_lmstudio_models(base_url: str) -> Optional[List[str]]:
+    """Fetch available models from LM Studio API.
+
+    Args:
+        base_url: LM Studio base URL (e.g., http://localhost:1234/v1)
+
+    Returns:
+        List of model IDs if successful, None if connection fails
+    """
+    try:
+        # Remove /v1 suffix if present and add /models endpoint
+        api_url = base_url.rstrip('/') + '/models'
+        response = requests.get(api_url, timeout=2)
+        response.raise_for_status()
+
+        data = response.json()
+        if 'data' in data and isinstance(data['data'], list):
+            models = [model['id'] for model in data['data'] if 'id' in model]
+            return models if models else None
+        return None
+    except (requests.RequestException, KeyError, ValueError) as e:
+        # Connection failed or invalid response
+        return None
+
+
 def select_shallow_thinking_agent(provider) -> str:
     """Select shallow thinking llm engine using an interactive selection."""
 
@@ -171,13 +198,78 @@ def select_shallow_thinking_agent(provider) -> str:
             ("GPT-OSS:latest (20B, local)", "gpt-oss:latest"),
             ("GLM-4.7-Flash:latest (30B, local)", "glm-4.7-flash:latest"),
         ],
+        "lm studio": [
+            ("Custom model (enter model name)", "custom"),
+        ],
     }
 
+    provider_key = provider.lower()
+
+    # For LM Studio, fetch models dynamically from API
+    if provider_key == "lm studio":
+        lmstudio_url = os.getenv("LMSTUDIO_BASE_URL", "http://localhost:1234/v1")
+
+        console.print(f"[dim]Fetching models from LM Studio at {lmstudio_url}...[/dim]")
+        available_models = fetch_lmstudio_models(lmstudio_url)
+
+        if available_models:
+            console.print(f"[green]✓ Found {len(available_models)} model(s)[/green]")
+
+            # Create choices from available models
+            choices = [questionary.Choice(model, value=model) for model in available_models]
+            # Add manual entry option at the end
+            choices.append(questionary.Choice("Enter model name manually", value="__manual__"))
+
+            choice = questionary.select(
+                "Select Your [Quick-Thinking LLM Engine]:",
+                choices=choices,
+                instruction="\n- Use arrow keys to navigate\n- Press Enter to select",
+                style=questionary.Style([
+                    ("selected", "fg:magenta noinherit"),
+                    ("highlighted", "fg:magenta noinherit"),
+                    ("pointer", "fg:magenta noinherit"),
+                ]),
+            ).ask()
+
+            if choice is None:
+                console.print("\n[red]No model selected. Exiting...[/red]")
+                exit(1)
+
+            # If manual entry selected, prompt for model name
+            if choice == "__manual__":
+                model_name = questionary.text(
+                    "Enter the model name:",
+                    instruction="\n- This should match the model ID from LM Studio",
+                ).ask()
+
+                if not model_name:
+                    console.print("\n[red]No model name entered. Exiting...[/red]")
+                    exit(1)
+
+                return model_name
+
+            return choice
+        else:
+            # Connection failed, show error and fall back to manual entry
+            console.print(f"[yellow]⚠ Could not connect to LM Studio at {lmstudio_url}[/yellow]")
+            console.print("[yellow]Please make sure LM Studio is running and a model is loaded.[/yellow]")
+
+            model_name = questionary.text(
+                "Enter the model name manually:",
+                instruction="\n- This should match the model ID from LM Studio",
+            ).ask()
+
+            if not model_name:
+                console.print("\n[red]No model name entered. Exiting...[/red]")
+                exit(1)
+
+            return model_name
+
     choice = questionary.select(
         "Select Your [Quick-Thinking LLM Engine]:",
         choices=[
             questionary.Choice(display, value=value)
-            for display, value in SHALLOW_AGENT_OPTIONS[provider.lower()]
+            for display, value in SHALLOW_AGENT_OPTIONS[provider_key]
         ],
         instruction="\n- Use arrow keys to navigate\n- Press Enter to select",
         style=questionary.Style(
@@ -238,13 +330,78 @@ def select_deep_thinking_agent(provider) -> str:
             ("GPT-OSS:latest (20B, local)", "gpt-oss:latest"),
             ("Qwen3:latest (8B, local)", "qwen3:latest"),
         ],
+        "lm studio": [
+            ("Custom model (enter model name)", "custom"),
+        ],
     }
 
+    provider_key = provider.lower()
+
+    # For LM Studio, fetch models dynamically from API
+    if provider_key == "lm studio":
+        lmstudio_url = os.getenv("LMSTUDIO_BASE_URL", "http://localhost:1234/v1")
+
+        console.print(f"[dim]Fetching models from LM Studio at {lmstudio_url}...[/dim]")
+        available_models = fetch_lmstudio_models(lmstudio_url)
+
+        if available_models:
+            console.print(f"[green]✓ Found {len(available_models)} model(s)[/green]")
+
+            # Create choices from available models
+            choices = [questionary.Choice(model, value=model) for model in available_models]
+            # Add manual entry option at the end
+            choices.append(questionary.Choice("Enter model name manually", value="__manual__"))
+
+            choice = questionary.select(
+                "Select Your [Deep-Thinking LLM Engine]:",
+                choices=choices,
+                instruction="\n- Use arrow keys to navigate\n- Press Enter to select",
+                style=questionary.Style([
+                    ("selected", "fg:magenta noinherit"),
+                    ("highlighted", "fg:magenta noinherit"),
+                    ("pointer", "fg:magenta noinherit"),
+                ]),
+            ).ask()
+
+            if choice is None:
+                console.print("\n[red]No model selected. Exiting...[/red]")
+                exit(1)
+
+            # If manual entry selected, prompt for model name
+            if choice == "__manual__":
+                model_name = questionary.text(
+                    "Enter the model name:",
+                    instruction="\n- This should match the model ID from LM Studio",
+                ).ask()
+
+                if not model_name:
+                    console.print("\n[red]No model name entered. Exiting...[/red]")
+                    exit(1)
+
+                return model_name
+
+            return choice
+        else:
+            # Connection failed, show error and fall back to manual entry
+            console.print(f"[yellow]⚠ Could not connect to LM Studio at {lmstudio_url}[/yellow]")
+            console.print("[yellow]Please make sure LM Studio is running and a model is loaded.[/yellow]")
+
+            model_name = questionary.text(
+                "Enter the model name manually:",
+                instruction="\n- This should match the model ID from LM Studio",
+            ).ask()
+
+            if not model_name:
+                console.print("\n[red]No model name entered. Exiting...[/red]")
+                exit(1)
+
+            return model_name
+
     choice = questionary.select(
         "Select Your [Deep-Thinking LLM Engine]:",
         choices=[
             questionary.Choice(display, value=value)
-            for display, value in DEEP_AGENT_OPTIONS[provider.lower()]
+            for display, value in DEEP_AGENT_OPTIONS[provider_key]
         ],
         instruction="\n- Use arrow keys to navigate\n- Press Enter to select",
         style=questionary.Style(
@@ -264,6 +421,9 @@ def select_deep_thinking_agent(provider) -> str:
 
 def select_llm_provider() -> tuple[str, str]:
     """Select the OpenAI api url using interactive selection."""
+    # Get LM Studio URL from environment or use default
+    lmstudio_url = os.getenv("LMSTUDIO_BASE_URL", "http://localhost:1234/v1")
+
     # Define OpenAI api options with their corresponding endpoints
     BASE_URLS = [
         ("OpenAI", "https://api.openai.com/v1"),
@@ -272,6 +432,7 @@ def select_llm_provider() -> tuple[str, str]:
         ("xAI", "https://api.x.ai/v1"),
         ("Openrouter", "https://openrouter.ai/api/v1"),
         ("Ollama", "http://localhost:11434/v1"),
+        ("LM Studio", lmstudio_url),
     ]
     
     choice = questionary.select(

tradingagents/llm_clients/factory.py

@@ -15,7 +15,7 @@ def create_llm_client(
     """Create an LLM client for the specified provider.
 
     Args:
-        provider: LLM provider (openai, anthropic, google, xai, ollama, openrouter)
+        provider: LLM provider (openai, anthropic, google, xai, ollama, openrouter, lm studio)
         model: Model name/identifier
         base_url: Optional base URL for API endpoint
         **kwargs: Additional provider-specific arguments
@@ -34,7 +34,7 @@ def create_llm_client(
     """
     provider_lower = provider.lower()
 
-    if provider_lower in ("openai", "ollama", "openrouter"):
+    if provider_lower in ("openai", "ollama", "openrouter", "lm studio"):
         return OpenAIClient(model, base_url, provider=provider_lower, **kwargs)
 
     if provider_lower == "xai":

tradingagents/llm_clients/openai_client.py

@@ -25,20 +25,22 @@ _PASSTHROUGH_KWARGS = (
 )
 
 # Provider base URLs and API key env vars
+# Format: (default_base_url, api_key_env_var)
 _PROVIDER_CONFIG = {
     "xai": ("https://api.x.ai/v1", "XAI_API_KEY"),
     "openrouter": ("https://openrouter.ai/api/v1", "OPENROUTER_API_KEY"),
     "ollama": ("http://localhost:11434/v1", None),
+    "lm studio": (None, "LMSTUDIO_API_KEY"),  # Base URL from LMSTUDIO_BASE_URL
 }
 
 
 class OpenAIClient(BaseLLMClient):
-    """Client for OpenAI, Ollama, OpenRouter, and xAI providers.
+    """Client for OpenAI, Ollama, OpenRouter, LM Studio, and xAI providers.
 
     For native OpenAI models, uses the Responses API (/v1/responses) which
     supports reasoning_effort with function tools across all model families
     (GPT-4.1, GPT-5). Third-party compatible providers (xAI, OpenRouter,
-    Ollama) use standard Chat Completions.
+    Ollama, LM Studio) use standard Chat Completions.
     """
 
     def __init__(
@@ -58,13 +60,27 @@ class OpenAIClient(BaseLLMClient):
         # Provider-specific base URL and auth
         if self.provider in _PROVIDER_CONFIG:
             base_url, api_key_env = _PROVIDER_CONFIG[self.provider]
-            llm_kwargs["base_url"] = base_url
+
+            # Special handling for LM Studio - get base URL from env
+            if self.provider == "lm studio":
+                base_url = os.environ.get("LMSTUDIO_BASE_URL", "http://localhost:1234/v1")
+
+            # Set base URL if available
+            if base_url:
+                llm_kwargs["base_url"] = base_url
+
+            # Handle API key
             if api_key_env:
                 api_key = os.environ.get(api_key_env)
                 if api_key:
                     llm_kwargs["api_key"] = api_key
+                else:
+                    # For LM Studio, use placeholder if no API key set
+                    if self.provider == "lm studio":
+                        llm_kwargs["api_key"] = "not-needed"
             else:
-                llm_kwargs["api_key"] = "ollama"
+                # For providers with no API key env var (like ollama)
+                llm_kwargs["api_key"] = "not-needed"
         elif self.base_url:
             llm_kwargs["base_url"] = self.base_url