Summary
The notes you already wrote become searchable context.
A years-deep Bear archive is mostly write-only in practice. Searching it by tag, by title and by the language of the code blocks inside is what makes an old note findable at the moment it would help.
What it is
A Python MCP server that reads the Bear app's notes on macOS. Search runs across full text, tags, titles and modification dates, and it parses fenced code blocks out of notes so snippets can be found by language.
What you get
- Full-text search across every note, and filtering by tag
- Title search, exact or partial
- Recent notes filtered by modification date
- Code blocks parsed out and categorised, so snippets are findable by language
- Kubernetes manifests located by what they are — deployment examples, service configs and the rest
- Word counts, content statistics and previews, with creation and modification dates
Requirements
macOS with Bear installed and notes in it, plus Python 3.8 or higher. The client launches the server from the checkout's virtual environment. No account, no key.