FAQ
Does ingest write wiki pages for me?
Section titled “Does ingest write wiki pages for me?”No — and that is deliberate. Capture stores what you saved, unchanged. Synthesis is a separate step done by your coding agent, following the synthesis contract; every page arrives as pending and joins the wiki only when you approve it. Every word that enters your knowledge base is one you can see and approve.
Do I need to be a programmer?
Section titled “Do I need to be a programmer?”You need a terminal for installing and running the commands, but you can delegate all of it — including running those commands — to an AI assistant. Most non-technical users live in the assistant workflow and never touch a script. Installation is one line: npm install -g @jerryjiao/knowflow.
Do I need an API key?
Section titled “Do I need an API key?”Only for semantic search, which is optional and off by default. Capture, graphs, health checks, and tag hubs all run without any key.
Which AI assistants work?
Section titled “Which AI assistants work?”Anything that can read and write files on your machine. ZCode, Codex CLI, Claude Code, and OpenCode each execute the same written synthesis contract; Cursor and friends work too. KnowFlow has no API lock-in — the contract and the conventions (templates, [[wikilinks]]) live in your files.
Is my data private?
Section titled “Is my data private?”Yes. KnowFlow runs entirely on your machine. The only network calls are fetching a URL you explicitly save — and, if you opt into semantic search, an embedding API. Both are visible in the code.
Can I use it with Obsidian?
Section titled “Can I use it with Obsidian?”Yes. The wiki/ folder is plain Markdown with [[wikilinks]] — open it as an Obsidian vault and everything just works. KnowFlow adds the parts Obsidian doesn’t have: capture, health checks, and generated graph data. (KnowFlow resolves links as paths from the wiki root, so write [[concepts/rag]], not [[RAG]].)
Why not just run RAG over my saved links?
Section titled “Why not just run RAG over my saved links?”RAG re-reads the raw pile from scratch for every question. KnowFlow compiles saves into durable, linked pages that accumulate over time — the cross-links and graph exist before you ever ask. The methodology page explains the difference.
Why Markdown files instead of a database?
Section titled “Why Markdown files instead of a database?”Because you should be able to leave. Files stay readable forever, version-control cleanly with Git, and can be processed by any tool — including AI assistants.
Is it on npm?
Section titled “Is it on npm?”Yes: npm install -g @jerryjiao/knowflow. The package is scoped because npm considers the bare name knowflow too similar to an existing package (know-flow). The CLI command is still just knowflow.
What does it cost?
Section titled “What does it cost?”Nothing. KnowFlow is MIT-licensed open source. The only possible cost is the embedding API if you enable semantic search.