Installation¶
Requirements¶
Golit requires Python 3.11 or newer. Installing it pulls in a small set of core dependencies:
| Dependency | Role |
|---|---|
| Litestar | The async server that hosts your app (Tier 1). |
| Polars | The columnar data engine; node values are Polars frames. |
| Lets-Plot | Grammar-of-graphics charts rendered to static SVG. |
| Uvicorn | The ASGI server golit run launches. |
The Rust reactive kernel ships precompiled in the wheel — you don't need a Rust toolchain to use Golit, only to build it from source.
Install¶
Verify it:
import golit
print(golit.__version__) # the Python package version
print(golit.kernel_version()) # the compiled Rust kernel version
Optional extras¶
Golit keeps the core install lean. Heavier or situational dependencies live behind extras you opt into — each is imported lazily, only when you actually use the feature.
Pulls in Plotly, Altair, and Bokeh. Return one of their figures from a view and Golit renders it as an interactive, client-side chart. (AnyChart needs no Python package — it loads from a CDN.) See Charts.
Pulls in DuckDB and PyArrow. Lets you write reactive nodes as SQL over your Polars frames with golit.sql().
Pulls in the redis client. Needed only when scaling to multiple workers, where server-side invalidations fan out across the fleet. See Deployment & scaling.
You can combine extras: pip install "golit[charts,sql,redis]".
Install from source¶
If you want to hack on Golit itself (including the Rust kernel), clone the repo and use the provided Makefile. It manages a uv-backed virtualenv and builds the extension with maturin:
git clone https://github.com/boadzie/golit
cd golit
make dev # uv venv (3.11) + deps + build the Rust kernel
make test # cargo test + pytest
See Contributing for the full developer workflow.
Next¶
You're ready to build. Head to Your first app.