Bear Notes MCP

Bear Notes MCP gives Claude direct read access to your Bear vault by querying the SQLite database that Bear stores locally on macOS. It's a single-purpose tool: you point it at Bear's database file, and Claude can then search notes by title, content, or tag, retrieve full note bodies, and list all tags in your vault. No API key, no sync service, just raw database queries. I'd reach for this when I want Claude to reference my personal notes during a conversation without manually copy-pasting. The tag exploration is particularly useful if you've built up a tagging system over years and want Claude to surface connections you've forgotten. For example, asking "What have I written about TypeScript performance?" and having Claude pull from notes tagged #typescript and #performance across years of entries is genuinely helpful. It also works well for migration prep: you can ask Claude to summarise note structures before moving to another system. The main quirk is that it's read-only. Claude can't create or edit notes, which is probably wise given Bear's proprietary format, but it does mean this is purely a retrieval tool. You'll need to know roughly where Bear stores its database (usually ~/Library/Group Containers/), and the server assumes you're running Bear 2.x on macOS. If you're on iOS or using Bear 1.x, this won't work. The SQLite approach also means no live sync: if you add a note in Bear, you might need to restart the MCP server for Claude to see it, depending on how Bear flushes writes. Performance is fine for vaults up to a few thousand notes. Beyond that, full-text searches might lag, but tag-based queries stay snappy. The codebase is small and readable, so if you want to tweak query logic or add filtering, it's straightforward. Skip this if you don't use Bear, obviously, or if you're on Windows or Linux. Also skip if you want Claude to write notes back: you'll need a different tool or a custom workflow with Bear's URL scheme.