TauricResearch/TradingAgents · error · ValueError

Unsupported LLM provider: {provider}

Error message

Unsupported LLM provider: {provider}

What it means

ValueError raised by create_llm_client (tradingagents/llm_clients/factory.py) when the configured provider string matches none of the known providers (openai, azure/openai-azure, bedrock, deepseek/other OpenAI-compatible names) and fails the is_openai_compatible check. It means the provider name in your LLM config is unrecognized.

Source

Thrown at tradingagents/llm_clients/factory.py:54

        return AnthropicClient(model, base_url, **kwargs)

    if provider_lower == "google":
        from .google_client import GoogleClient
        return GoogleClient(model, base_url, **kwargs)

    if provider_lower == "azure":
        from .azure_client import AzureOpenAIClient
        return AzureOpenAIClient(model, base_url, **kwargs)

    if provider_lower == "bedrock":
        from .bedrock_client import BedrockClient
        return BedrockClient(model, base_url, **kwargs)

    from .openai_client import OpenAIClient, is_openai_compatible
    if is_openai_compatible(provider_lower):
        return OpenAIClient(model, base_url, provider=provider_lower, **kwargs)

    raise ValueError(f"Unsupported LLM provider: {provider}")

View on GitHub (pinned to a33fd4c0f1)

Solutions

1. Use a supported provider string: 'openai', 'azure', 'bedrock', or an OpenAI-compatible name recognized by is_openai_compatible (e.g. 'deepseek', 'openai-compatible' base URLs).
2. If the vendor is OpenAI-API-compatible, point provider at the openai path and set base_url to the vendor endpoint.
3. Upgrade the package — new providers are added over time — and check factory.py for the current accepted list.

Example fix

# before
config['llm_provider_quick'] = 'gcp'

# after
config['llm_provider_quick'] = 'openai'
config['quick_think_llm_base_url'] = 'https://your-openai-compatible-endpoint/v1'

SUPPORTED_PROVIDERS = {'openai', 'azure', 'bedrock'} | {
    # plus whatever is_openai_compatible accepts; check factory.py for your version
}

def is_supported_provider(provider: str) -> bool:
    p = provider.strip().lower()
    return p in SUPPORTED_PROVIDERS or p.replace('-', '').replace('_', '') in SUPPORTED_PROVIDERS

from tradingagents.llm_clients.factory import create_llm_client

try:
    client = create_llm_client(model, base_url, provider=provider)
except ValueError:
    client = create_llm_client(model, vendor_base_url, provider='openai')  # OpenAI-compatible fallback

When it happens

Trigger: Setting the quick-think/deep-think LLM provider in DEFAULT_CONFIG (or TradingAgentsGraph config) to a string like 'myprovider', 'google', or 'Azure' variants that don't match factory branches. The lowercased value falls through every branch and the final raise fires.

Common situations: Typos in provider names ('open_ai', 'awzure'); assuming a provider is supported because langchain supports it; missing integration for a vendor added after your package version.

Related errors

• expected a boolean ({'/'.join(_BOOL_TRUE + _BOOL_FALSE)}), g
• Invalid value for {env_var}: {exc}
• unknown analyst key: {analyst_key}
• at least one analyst must be selected
• llm_max_retries must be an integer, not a boolean: {value!r}

AI-assisted analysis of TauricResearch/TradingAgents@a33fd4c0f1 (2026-08-14). Data as JSON: /api/errors/d087f8b7b658ea4d. Report an issue: GitHub.