What one project knows about itself
Every new session starts from zero. Give it what the project already knows.
A session re-reads the repo and guesses the architecture. Knowledge keeps what one project knows about itself in one place: its memory files in one editor, an architecture map that says how stale it is, and code graphs agents can query instead of grepping.
Instructions and memory, in one editor
Project instructions, auto-memory and Serena notes in one editor. Edit, preview, save: every save keeps a backup. A stale-fact lint catches memory that disagrees with the repo, like a PR the memory calls open when git says it merged.
- A conflict guard on every save
- Finished entries move out of the index agents always load
- Stale facts flagged against git
A map of the repo, with a staleness badge
An architecture map of the repo: modules by layer (UI, domain, data, server), and flows you can trace step by step, file by file. The map says how stale it is, per commit.
- Machine-readable layers, modules and flows
- Pick a flow and follow every step
- Freshness stamped per commit
Real symbols, not grep
Serena gives agents a language server, and the dashboard shows which tools they call. Graphify turns the code into a graph of communities: click any symbol to see what it touches.
- Symbol-level search and references
- A graph of communities
- What a symbol touches, one click away
screenshots
The knowledge pages, up close.
Click any screenshot to open it full size. The projects are fictional.
Run your agents like a fleet.
Free on GitHub. One binary on your machine, no cloud, no account. Your sessions show up the moment it starts.