crewAIInc/crewAI · error · FileNotFoundError

No JSON crew definition configured in [tool.crewai].definiti

Error message

No JSON crew definition configured in [tool.crewai].definition

What it means

A FileNotFoundError raised when running a JSON crew with no explicit crew path: the `crew_path` argument was None and configured_project_json_crew() also returned None, meaning pyproject.toml has no `[tool.crewai].definition` entry pointing at a crew.json(c) file. Because it is raised (not caught) inside the CLI, it propagates as a traceback rather than a friendly message.

Source

Thrown at lib/cli/src/crewai_cli/run_crew.py:327

    crew_path: str | Path | None = None,
    inputs: str | None = None,
) -> Any:
    """Load and run a JSON-defined crew."""
    from dotenv import load_dotenv

    env_file = Path.cwd() / ".env"
    if env_file.exists():
        load_dotenv(env_file, override=True)

    # JSON crews run in-process, so export the trained-agents file directly
    # instead of forwarding it to a subprocess like classic crews do.
    if trained_agents_file:
        os.environ[CREWAI_TRAINED_AGENTS_FILE_ENV] = trained_agents_file

    if crew_path is None:
        crew_path = configured_project_json_crew()
    if crew_path is None:
        raise FileNotFoundError(
            "No JSON crew definition configured in [tool.crewai].definition"
        )
    crew_path = Path(crew_path)

    provided = parse_inputs_json(inputs)

    if is_dmn_mode_enabled():
        return _run_json_crew_without_tui(crew_path, provided)

    crew_run_app_cls, crew, default_inputs, task_names, agent_names = (
        _load_json_crew_for_tui(crew_path)
    )
    runtime_inputs = _resolve_crew_inputs(
        crew, default_inputs, provided, interactive=is_interactive()
    )

    app = crew_run_app_cls(
        crew_name=crew.name or "Crew",

View on GitHub (pinned to 754d7323be)

Solutions

1. Add the definition entry to pyproject.toml: `[[tool.crewai]]` with `definition = "crew.json"` (path to your JSON crew).
2. Or pass the path explicitly if the command supports it, so discovery is not needed.
3. Run the command from the project root where pyproject.toml lives.
4. If you meant to run a classic Python crew, use the classic flow entry point instead of the JSON-crew path.

Example fix

# before
# pyproject.toml has no [tool.crewai] table
$ crewai run
# FileNotFoundError: No JSON crew definition configured in [tool.crewai].definition

# after
# pyproject.toml
[tool.crewai]
definition = "crew.json"

from pathlib import Path
import tomllib

pp = Path("pyproject.toml")
assert pp.is_file(), "run from project root"
data = tomllib.loads(pp.read_text())
definition = data.get("tool", {}).get("crewai", {}).get("definition")
if not definition or not (Path.cwd() / definition).is_file():
    raise SystemExit("[tool.crewai].definition missing or points at a nonexistent file")

When it happens

Trigger: Invoking `crewai run` (JSON-crew path) in a directory whose pyproject.toml lacks `[[tool.crewai]] definition = "crew.json"`; running from the wrong working directory so pyproject.toml isn't found; a scaffolded project that uses classic Python crews (crew.py) being invoked on the JSON path.

Common situations: Running the CLI from a subdirectory instead of the project root; a pyproject.toml edited by hand that dropped the `[tool.crewai]` table; mixing up classic (crew.py) and declarative JSON crew workflows.

Related errors

• project_name is required to create a ZIP deployment
• An unexpected error occurred while running the JSON crew: {e
• Invalid --definition path: {definition} does not exist.
• Error: Unknown or readonly configuration key '{key}'
• No organisation set. Run `crewai org switch <org_id>` first,

AI-assisted analysis of crewAIInc/crewAI@754d7323be (2026-08-15). Data as JSON: /api/errors/7c8b5caa469b6f84. Report an issue: GitHub.