Writing documentation locally & syncing

Your project documentation lives as Markdown files in the connected repository. Lalabase mirrors it read-only — the repository stays the single source of truth.

This is by design: documentation isn't written inside Lalabase, it's maintained by the developers with repository access in their normal workflow. That brings two benefits. First, there's exactly one source of truth: if a team ever stops using Lalabase, the docs aren't lost — they remain as Markdown in their own repository. Second, colleagues without repository access — non-technical team members, for example — still read the always-current documentation right inside Lalabase, with no Git or repo access of their own.

Some teams also have their own AI tool summarise changes in the repository before committing — that's purely a workflow tip outside Lalabase, not a Lalabase feature.

Writing documentation

1. Create a docs folder in your repository — the default is .lalabase/doku/.
2. Write the pages as Markdown files and commit them.
3. Push the changes to GitHub as usual.

Connecting it in Lalabase

1. First connect the repository to the project.
2. Open the Documentation area and click Connect documentation.
3. Pick the repository, enter the folder (default .lalabase/doku/), and give it a name.
4. Save, then run the first sync via Sync now.

Kept up to date automatically

Once the repository is connected with a push webhook, the docs here update on their own — nobody has to copy anything over by hand. If you like, let an AI tool in the repository help write and maintain the documentation while coding. To Lalabase that's transparent: it simply mirrors the latest state from the repository.

Search and chat

The mirrored docs aren't an isolated corner. Their pages appear in global search, and the project chat can draw on them. So you can search them like anything else, or ask the chat a question about the documentation and get an answer based on the current pages.