assafelovic/gpt-researcher · error · ValueError

Unsupported type {type_hint} for key {key}

Error message

Unsupported type {type_hint} for key {key}

What it means

convert_env_value handles bool, int, float, str, list, dict and None; any other type hint (or a value that doesn't fit the str branch) falls to the final else and raises 'Unsupported type ... for key ...'.

Source

Thrown at gpt_researcher/config/config.py:306

            # Env values are often hand-edited (trailing commas, single quotes).
            # Bare `list` has get_origin(None); typing.List[...] has origin list.
            try:
                value = json_repair.loads(env_value)
            except Exception as exc:
                raise ValueError(f"Cannot convert {env_value} to list") from exc
            if not isinstance(value, list):
                raise ValueError(f"Cannot convert {env_value} to list")
            return value
        elif type_hint is dict:
            try:
                value = json_repair.loads(env_value)
            except Exception as exc:
                raise ValueError(f"Cannot convert {env_value} to dict") from exc
            if not isinstance(value, dict):
                raise ValueError(f"Cannot convert {env_value} to dict")
            return value
        else:
            raise ValueError(f"Unsupported type {type_hint} for key {key}")


    def set_verbose(self, verbose: bool) -> None:
        """Set the verbosity level."""
        self.llm_kwargs["verbose"] = verbose

    def get_mcp_server_config(self, name: str) -> dict:
        """
        Get the configuration for an MCP server.
        
        Args:
            name (str): The name of the MCP server to get the config for.
                
        Returns:
            dict: The server configuration, or an empty dict if the server is not found.
        """
        if not name or not self.mcp_servers:
            return {}

View on GitHub (pinned to 6f998577d5)

Solutions

1. Change the attribute's annotation to a supported type (str/int/float/bool/list/dict)
2. Set the value programmatically instead of via env var
3. Report/patch convert_env_value to support the type

Example fix

# before
my_field: set[str]  # env conversion raises
# after
my_field: list[str]

SUPPORTED = (bool, int, float, str, list, dict, type(None))
# when subclassing/patching Config, check annotations:
for k, t in Config.__annotations__.items():
    assert t in SUPPORTED or str(t).startswith('typing'), f"unsupported {k}: {t}"

try:
    cfg = Config()
except ValueError as e:
    if "Unsupported type" in str(e):
        # set programmatically instead of env
        cfg = Config(); setattr(cfg, key, value)
    else: raise

When it happens

Trigger: A Config attribute annotated with an exotic type (set, custom class, tuple) whose matching env var is set, so conversion is attempted; typically surfaces after library changes introduce new annotations.

Common situations: Version drift: your fork adds a typed attribute without updating convert_env_value; passing nested typed structures via env.

Related errors

• Cannot convert {env_value} to any of {args}
• Embedding provider not found.
• Invalid retriever(s) found: {', '.join(invalid_retrievers)}.
• Set SMART_LLM or FAST_LLM = '<llm_provider>:<llm_model>' Eg
• Invalid reasoning effort: {reasoning_effort_str}. Valid opti

AI-assisted analysis of assafelovic/gpt-researcher@6f998577d5 (2026-08-28). Data as JSON: /api/errors/f3aa6f433f55c38d. Report an issue: GitHub.