Install
Run this in a terminal on macOS (Apple Silicon) or Linux (x64):~/.aius and adds ~/.aius/bin to
your PATH. Open a new terminal and check it:
aius command is already in ~/.aius/bin, the installer keeps it as
aius.backup-… next to the new one.
Update by running aius upgrade, or the install command again. Install a specific version,
for example to go back to the one you had, by setting AIUS_VERSION:
~/.aius/bin/aius and ~/.aius/releases, and the
# AIUS lines the installer added to your shell profile (~/.zshrc,
~/.bashrc or ~/.bash_profile). Your sessions, sign-in and settings are in
~/.local/share/aius-agent, ~/.local/state/aius-agent, ~/.cache/aius-agent
and ~/.config/aius-agent; delete those folders too to remove them.
Sign in
The firstaius start asks you to sign in (see Quickstart).
On a server without a browser, sign in with a code instead and approve it from
any other device:
aius platform status prints the account the terminal is signed in to.
aius platform logout signs this terminal out and revokes its access; your
browser session is separate.
Python for data work
The data tools run Python on your computer, in the environment of the project you are working in. AIUS looks for an interpreter in this order: the project’s.venv, then an active virtual environment (VIRTUAL_ENV), then an active
conda environment (CONDA_PREFIX), then python3.
A project-local environment keeps each analysis reproducible:
| Capability | Packages |
|---|---|
| Profiling, cleaning, modelling | pandas, numpy, scikit-learn |
| SQL over CSV, Parquet, Excel and DuckDB files | duckdb, pyarrow |
| Figures | matplotlib |
| Notebooks | nbformat, nbclient, ipykernel |
python_environment tool reports the
interpreter it selected and the packages it found.