wiki

You can store persistent Markdown project documentation across sessions by saving pages as files in a repo-backed knowledge base. This approach supports keyword and tag search, ensuring architecture decisions and troubleshooting notes survive between turns.

Keyword and tag search for a project knowledge base works by querying indexed Markdown pages using specific terms and categories without requiring embeddings. It enables fast retrieval of architecture, debugging, and decision notes based on assigned tags.

The best way to maintain a lifecycle log for project decisions is using a wiki system that enforces an index and log. This captures ongoing project memory and evidence, allowing you to review and refresh knowledge entries over time.

Yes, you can use cross-reference links to connect related architecture notes by inserting wiki-style page links. This connects related concepts and decisions within your Markdown knowledge base, building a navigable network of project documentation.

No, this knowledge base tool does not require embeddings for searching documentation. It enforces keyword and tag search to query pages, with an optional legacy fallback, allowing text retrieval without vector database dependencies.

You ingest an architecture page into a Markdown wiki by running an ingest command with a JSON input containing the title, content, tags, and category. The system then stores the page in repo-backed storage for persistent access.