The engine
pypilot is a single native binary, written in Rust, that is both a command-line tool and a language server. The Zed extension, the VS Code extension and the pypilot-cli package all run it; nothing else holds logic.
text
pypilot <MODE> [ARG] [--path <dir>]| Mode | Changes things? | Does |
|---|---|---|
doctor | no | Environment and compatibility report (the default mode); --fix 1.3 then runs setup |
check <package> | no | Can this package run on this project's Python? |
explain [package] 1.3 | no | Why these Pythons: requires_python vs wheels per platform |
setup | yes | Build the environment on the recommended Python |
install <package> | yes | Install one package and record it in the manifest |
fix python | yes | Rebuild the environment on the right Python |
fix cuda | yes | Re-pin torch/tensorflow to the build for your driver |
update-data | cache only | Refresh the bundled data tables now; --check 1.3 only compares |
migrate-conda | writes a file | environment.yml to pyproject.toml (or pixi.toml 1.3) |
migrate-to-pixi 1.3 | with --write | Preview or write a pixi.toml |
install-pixi 1.3 | installs pixi | Run pixi's official installer |
logs 1.3 | no | The local log's path, or its last entries |
completions <shell> 1.3 | no | Shell completion script |
version 1.3 | no | Version, commit, target and data-table state |
lsp | none | Language server over stdio |
Modes marked 1.3 are new in engine 1.3.0; older releases have the others. doctor, check and explain only read. setup, fix and install share their code with the matching editor actions, so the terminal and the editor cannot disagree, and 1.3 each takes --dry-run to show its plan first.
$ pypilot doctor
PyPilot doctor: C:\src\gpu
package manager : Auto
── Environment ──────────────────────────────────────────
uv : present (v0.13.5, on PATH)
virtualenv : none
interpreters : 3.15 (python3), 3.14 (C:\Python314\python.exe)
data tables : refreshed 3 hours ago (next in 6 days); 1 of 5 bundled
── Project ──────────────────────────────────────────────
files : pyproject.toml
requires-python : >=3.10
dependencies : torch, transformers
── Compatibility ────────────────────────────────────────
supported Python: 3.11–3.15
recommended : Python 3.15
• torch 3.11–3.15
• transformers 3.11–3.15
── Findings ─────────────────────────────────────────────
[WARN ] No environment set up yet
Dependencies support Python 3.11–3.15. Recommended: 3.15 (upper bound set by torch, which supports 3.11–3.15).
[INFO ] torch build matched to your driver: cu132
NVIDIA RTX A5000 Laptop GPU's driver (596.86) supports CUDA up to 13.2. NVIDIA RTX A5000 Laptop GPU (Ampere, sm_86) needs a build against CUDA 11.1 or newer. Recommended: pip install torch==2.15.0 --index-url https://download.pytorch.org/whl/cu132
Everyday use
bash
pypilot # same as pypilot doctor
pypilot doctor --path ../api # another folder
pypilot check torch # will torch run here?
pypilot check "numpy<2" # requirements work too
pypilot explain # what blocks each Python in this project
pypilot setup # build .venv (or the pixi env) on the right Python
pypilot install httpx # add a dependency through uv, pip or pixi
pypilot setup --dry-run # (1.3) what setup would run, changing nothing
pypilot doctor --json # (1.3) the whole report for a script or an issueThe full reference with every flag is on Command reference. For scripting, see Machine-readable output; for previews, Dry runs; for the local log and tab completion, Logs, version & completions.
Where things live
| What | Where |
|---|---|
| Global config | config.toml in the platform config directory (for example ~/.config/pypilot/config.toml on Linux) |
| Project config | .zed/pypilot.toml, else pypilot.toml, in the project root |
| PyPI metadata cache, refreshed data tables, the local log 1.3 | the platform cache directory: %LOCALAPPDATA%\PyPilot\pypilot\cache on Windows, ~/Library/Caches/dev.PyPilot.pypilot on macOS, ~/.cache/pypilot on Linux (the log is in logs/) |
See Configuration.
