Installation¶

plane-agent is a standard Python package and a prebuilt container image. Pick the path that matches how you want to run it.

Requirements¶

Python 3.11 – 3.14.
A reachable Plane workspace — either Plane Cloud or a self-hosted instance (see Backing Platform to deploy one locally), plus a PLANE_API_KEY and PLANE_WORKSPACE_SLUG.

From PyPI (recommended)¶

pip install plane-agent

The base install already pulls in agent-utilities[agent,logfire], so both the MCP server (plane-mcp) and the A2A agent (plane-agent) console scripts are available immediately, with Logfire tracing enabled.

Optional extras¶

Extra	Install	Pulls in
`test`	`pip install "plane-agent[test]"`	`pytest`, `pytest-asyncio`, `pytest-cov`, `pytest-xdist` for the test suite

# Typical: run the MCP server and the agent
pip install plane-agent

From source¶

git clone https://github.com/Knuckles-Team/plane-agent.git
cd plane-agent
pip install -e ".[test]"          # editable install with the test extra

With uv:

uv pip install -e ".[test]"
uv run plane-mcp

Prebuilt Docker image¶

A multi-stage, slim image is published on every release (entrypoint plane-mcp):

docker pull knucklessg1/plane-agent:latest

docker run --rm -i \
  -e PLANE_BASE_URL=https://api.plane.so \
  -e PLANE_API_KEY=your_plane_api_key \
  -e PLANE_WORKSPACE_SLUG=your-workspace \
  knucklessg1/plane-agent:latest        # stdio transport (default)

For an HTTP server with a published port and the agent server, see Deployment.

Verify the install¶

plane-mcp --help
plane-agent --help
python -c "import plane_agent; print(plane_agent.__version__)"

Next steps¶

Deployment — run it as a long-lived MCP server and agent behind Caddy + DNS.
Usage — call the tools, the API, and the CLI.
Configuration — every environment variable.