An AI-native, self-updating Linux distribution: local models as a system service, per-app durable context, an append-only Ledger, and an OpenAI-compatible endpoint right on the machine. Flash it, boot it, build against it.
Good first issues, straight from the repo. Pick one, comment to claim it, and open a PR.
Sign in with GitHub to claim issues and track your open PRs here.
Sign in with GitHubFlash the latest USB image and boot it — it runs from the stick, then installs to disk when you're ready. Updates are A/B with automatic rollback on a bad boot.
# 1. Download lisa-usb-<version>.raw.zst from Releases, then:
zstd -d lisa-usb-*.raw.zst -o lisa.raw
sudo dd if=lisa.raw of=/dev/<your-usb> bs=4M status=progress oflag=sync
# 2. Boot it. To install onto the internal disk (erases it):
lisa install /dev/<internal-disk>
# 3. Update in place later (A/B, auto-rollback on a bad boot):
lisa update --rebootIntelligence is an OpenAI-compatible endpoint on the machine — build against it with zero Lisa-specific dependencies. Guided generation means typed output that always parses.
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:7777/v1", api_key="local")
r = client.chat.completions.create(
model="lisa",
messages=[{"role": "user", "content": "Extract the recipe.\n\n" + text}],
response_format={"type": "json_schema",
"json_schema": {"name": "recipe", "schema": SCHEMA}})
# always valid JSON for SCHEMAThe design is the source of truth, and every non-obvious decision is written down.