Skip to content

Installation and upgrades

uvx hailer ... fetches Hailer from PyPI the first time, keeps it in uv's cache with an environment of its own (marimo, Polars, DuckDB and the rest), and runs it from there. No pyproject.toml or .venv is needed in your analysis folder.

To use the latest release, run uvx hailer@latest. Use uvx hailer@X.Y.Z to select a specific version.

For a permanent hailer command:

uv tool install hailer

Then use hailer wherever these instructions say uvx hailer. Upgrade that installation with uv tool upgrade hailer. You can also use pip install hailer in your own Python 3.12+ environment. See uv's tools guide for more installation options.

Requirements

  • Windows 11 is the primary target; macOS and Linux work too.
  • Hailer needs Python 3.12 or newer; uv can download it automatically.
  • Keep the notebook open in a web browser while using Hailer.
  • Every model provider needs an API key (see Model configuration).
  • Docker, installed and running: notebook code runs in a Docker container by default. Install Docker Desktop on Windows or macOS (on Windows, keep it on Linux containers, its default) or Docker Engine on Linux. The first start downloads Hailer's kernel image (uvx hailer kernel pull does it ahead of time). See Isolated kernel (Docker).
  • Without Docker, Hailer can run notebook code on your machine, as you, with your files and network, if you choose that yourself: runtime = "unsafe-local" under [kernel] in hailer.toml. See Kernel runtimes before you do.

For source checkouts and sample data generation, see Development.