harry0703/MoneyPrinterTurbo · error · ValueError

{llm_provider}: base_url is not set, please set it in the co

Error message

{llm_provider}: base_url is not set, please set it in the config.toml file.

What it means

Base-URL guard in _generate_response: the provider declares requires_base_url (typical for OpenAI-compatible custom endpoints like azure or self-hosted gateways) but the resolved base_url is empty. The SDK would otherwise default to api.openai.com and send the request (and the key) to the wrong service.

Source

Thrown at app/services/llm.py:205

        extra_values = {
            field.config_suffix: (
                runtime_app_config.get(provider.config_key(field.config_suffix), "")
                or field.default_value
            )
            for field in provider.extra_fields
        }

        if provider.requires_api_key and not api_key:
            raise ValueError(
                f"{llm_provider}: api_key is not set, please set it in the config.toml file."
            )
        if provider.requires_model_name and not model_name:
            raise ValueError(
                f"{llm_provider}: model_name is not set, please set it in the config.toml file."
            )
        if provider.requires_base_url and not base_url:
            raise ValueError(
                f"{llm_provider}: base_url is not set, please set it in the config.toml file."
            )

        for field in provider.extra_fields:
            if field.required and not extra_values[field.config_suffix]:
                raise ValueError(
                    f"{llm_provider}: {field.config_suffix} is not set, "
                    "please set it in the config.toml file."
                )

        if adapter == "qwen":
            import dashscope
            from dashscope.api_entities.dashscope_response import GenerationResponse

            dashscope.api_key = api_key
            response = dashscope.Generation.call(
                model=model_name, messages=[{"role": "user", "content": prompt}]
            )

View on GitHub (pinned to 1f9f19c202)

Solutions

1. Set <provider>.base_url in config.toml to the endpoint root (e.g. http://localhost:11434/v1) including the /v1 suffix the SDK expects
2. Verify reachability of the URL from the worker process (curl) to confirm the right host/port
3. Keep credentials out of the URL where possible; the code sanitizes userinfo in errors, but not sending them at all is safer
4. Restart/reload the service so the new config takes effect for new tasks

Example fix

# config.toml before
[openai-compat]
api_key = "sk-local"  # base_url missing -> raises

# config.toml after
[openai-compat]
api_key = "sk-local"
base_url = "http://127.0.0.1:8000/v1"

base_url = cfg.get(provider.config_key("base_url"), "")
if provider.requires_base_url and not base_url:
    raise ConfigError(f"{provider.config_key('base_url')} is empty")

When it happens

Trigger: Providers whose registry entry sets requires_base_url=true where <provider>.base_url is missing or empty in the effective config — e.g. configuring an azure/openai-compatible provider without its endpoint URL.

Common situations: Switching llm_provider to a self-hosted provider (one-api, vllm, ollama-openai bridges) without setting base_url, TOML section renamed so the key is ignored, or the field filled only in the WebUI while the worker reads config.toml.

Related errors

• [{llm_provider}] returned empty choices
• [{llm_provider}] returned empty message
• {llm_provider}: unsupported llm provider
• {llm_provider}: api_key is not set, please set it in the con
• {llm_provider}: model_name is not set, please set it in the

AI-assisted analysis of harry0703/MoneyPrinterTurbo@1f9f19c202 (2026-08-14). Data as JSON: /api/errors/5ad9b6a0faad2c08. Report an issue: GitHub.