What is kknow

kknow is a fast, local-first knowledge base for plain Markdown files. You write notes in a desktop window, and the files stay on your disk as ordinary .md files.

Three ideas

Files are the source of truth

A vault is an ordinary folder of .md files. You can open the same folder in any other editor, keep it in git or put it in Dropbox. kknow does not add a database next to your notes.

kknow keeps other data, but all of it is a cache:

• the link graph and the backlinks;
• the full-text search index;
• the state of the last sync.

kknow can build the cache again at any time. If you remove it, you lose no notes.

One core, many frontends

The logic is in Rust: parsing, links, search, sync and file watching. The same core serves three frontends:

• the window — an editor with live preview, built on Tauri;
• the CLI (command-line interface) — the kknow program in the terminal;
• the MCP server — the same vaults for an AI agent, through the Model Context Protocol.

A note that you save in the window is visible at once in kknow search and to the agent.

Local first

kknow works without a network and without an account. Sync is optional, and you select the storage: an S3 bucket, a MinIO server, iCloud Drive or a plain folder. kknow never loads images or themes from the web, so a note from another person cannot tell a third-party server that you opened it.

Who it is for

kknow is a good fit if you:

• want an Obsidian-like editor, but on plain files that other tools can use;
• keep several separate folders of notes, for example work, study and personal;
• use the terminal or AI agents with your notes;
• want to own the sync storage.

What kknow is not

• It is not a cloud service. You do not get hosting or a web version.
• It does not have plugins in JavaScript. You extend it with custom commands in any language.
• It does not sync through git yet. You can still keep a vault in a git repository: kknow shows the changes from git pull as any outside edit.

Next steps

• Install kknow and open your first vault.
• Read the feature overview.
• Read how it works inside: Architecture.