Skip to main content
Install the runtime from source. Nothing is published to PyPI yet, so you clone the three packages side by side and let uv resolve them from the sibling checkouts.

​
Requirements

  • Python 3.13 or newer.
  • uv to resolve and sync dependencies.
  • Redis 7.0+ and Postgres only for the features that need them — access control, connectors, and interactions. Local tools and mounted MCP servers need no services at all.

​
Install from source

tai42-skeleton’s [tool.uv.sources] points at ../tai-contract and ../tai-kit, so the three repositories must sit next to each other. Clone them, then sync the skeleton:
Once the packages are published, pip install tai42-skeleton replaces the from-source flow. Until then, install from the side-by-side checkouts.

​
Add the standard toolbox

The toolbox extra pulls in the tai42-toolbox contrib package — its generic tools and its composition extensions (chain, batch) load straight from a manifest:
Once the packages are published, pip install tai42-skeleton[toolbox] works too. See the standard toolbox guide for what ships in it.

​
The tai command

Installing the package puts the entrypoint on your path under two names, tai and tai42. Both run the same command tree — tai is the documented one, and tai42 is an alias for environments where an unrelated tai already sits on your PATH. Three subcommands run the long-lived processes:
  • tai serve — run the MCP server.
  • tai backend — run the agent / worker backend process.
  • tai metrics — serve the Prometheus metrics endpoint.
Run tai --help for the full command tree, or browse the CLI reference. Next, boot a server and call a tool in the quickstart.