langchain-ai/deepagents · error · ValueError

Server '{server_name}' has type '{server_type}' (remote) but

Error message

Server '{server_name}' has type '{server_type}' (remote) but also declares a 'command' field. Remove 'command' or set `"type": "stdio"`.

What it means

A server whose resolved type is remote must not declare a 'command' field; that field is exclusive to stdio servers. The error tells the user to remove 'command' or switch the type to stdio.

Source

Thrown at libs/code/deepagents_code/mcp_tools.py:893

        raise TypeError(error_msg)

    server_type = _resolve_server_type(server_config)

    if server_type in _SUPPORTED_REMOTE_TYPES:
        if "url" not in server_config:
            error_msg = (
                f"Server '{server_name}' with type '{server_type}' "
                "missing required 'url' field"
            )
            raise ValueError(error_msg)

        if "command" in server_config:
            error_msg = (
                f"Server '{server_name}' has type '{server_type}' (remote) "
                "but also declares a 'command' field. Remove 'command' or "
                'set `"type": "stdio"`.'
            )
            raise ValueError(error_msg)

        headers = server_config.get("headers")
        if headers is not None and not isinstance(headers, dict):
            error_msg = f"Server '{server_name}' 'headers' must be a dictionary"
            raise TypeError(error_msg)

        if isinstance(headers, dict):
            for name, value in headers.items():
                if not isinstance(value, str):
                    error_msg = (
                        f"Server '{server_name}' header {name!r} must be "
                        f"a string, got {type(value).__name__}"
                    )
                    raise TypeError(error_msg)
    elif server_type == "stdio":
        if "command" not in server_config:
            error_msg = f"Server '{server_name}' missing required 'command' field"
            raise ValueError(error_msg)

View on GitHub (pinned to a1af029e6e)

Solutions

1. Remove the 'command' (and 'args') fields from the remote server entry
2. Set "type": "stdio" instead if the server is launched locally via command
3. Re-run validation after editing to confirm the config is clean

Example fix

// before
{"api": {"type": "http", "url": "https://mcp.example.com", "command": "npx"}}
// after
{"api": {"type": "http", "url": "https://mcp.example.com"}}

def remote_entry_clean(cfg: dict) -> bool:
    if cfg.get('type') in ('http', 'sse', 'streamable-http'):
        return 'command' not in cfg
    return True

try:
    tools = await resolve_and_load_mcp_tools(config)
except ValueError as e:
    if "declares a 'command' field" in str(e):
        log.error('remote server must not define command: %s', e)
    raise

When it happens

Trigger: Config mixing remote fields (type: http/sse + url) with a leftover 'command' key; converting a stdio entry to remote without deleting 'command'.

Common situations: Editing an existing stdio entry to point at a remote URL while keeping the launch command; merging two config files that define the same server differently.

Related errors

• Invalid MCP server name {server_name!r}: token storage names
• MCPServerInfo {self.name!r}: status={self.status!r} cannot c
• Server '{server_name}' config must be a dictionary
• Server '{server_name}' with type '{server_type}' missing req
• {what} must be absolute: {path}

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/ae9b0c7ab86c0a15. Report an issue: GitHub.