MCP Bear
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@MCP Bear search for my notes about project ideas"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP Bear
A Python-based Model Context Protocol (MCP) server that provides access to Bear Notes.
Bear stores notes in a SQLite database. This MCP server provides both read and write access to your notes. See: https://bear.app/faq/where-are-bears-notes-located
Features
Read Operations
Read all notes (active and archived) from Bear's SQLite database
Search notes by text content, tag, or unique ID
List all tags
Get notes by specific tag
Get archived notes separately
Note Management
Create new notes with title, content, and tags
Append, prepend, or replace text in existing notes
Archive/unarchive notes for organization
Move notes to trash
Tag Management
Add tags to existing notes
Rename tags across all notes
Open Bear showing notes with specific tag
Bear Integration
Open specific notes in Bear
Search within Bear app
All operations integrate seamlessly with Bear's UI
Write operations use Bear's official x-callback-url API for safe, reliable modifications.
Related MCP server: Bear App MCP Server
Requirements
Python 3.10 or higher (tested with 3.14)
Bear note application (macOS)
Access to Bear database
Installation
Development Setup
# Clone the repository
git clone https://github.com/maxim-ist/mcp-bear.git
cd mcp-bear
# Install dependencies (using mise for Python version management)
mise install
mise exec -- python -m venv .venv
.venv/bin/pip install -e .Install via pip (when published)
pip install mcp-bearClaude Desktop Configuration
Add this to your claude_desktop_config.json:
Using Python directly
{
"mcpServers": {
"bear": {
"command": "python",
"args": [
"-m",
"mcp_bear.server"
],
"env": {
"PYTHONPATH": "/Users/YOUR_USERNAME/Work/mcp-bear/src"
}
}
}
}Using installed package
{
"mcpServers": {
"bear": {
"command": "mcp-bear"
}
}
}Available Tools
When the server is started, the following MCP tools become available:
Read Operations (Basic)
get_notes: Retrieves all non-archived notesget_tags: Lists all tagsget_notes_like: Searches for notes containing specific text
Read Operations (Advanced)
get_note_by_id: Get a specific note by its unique identifierget_notes_by_tag: Get all notes with a specific tagget_archived_notes: Get all archived notes
Note Management
create_note: Create a new note with optional title, text, tags, and pin statusadd_text: Add text to an existing note (append, prepend, or replace)trash_note: Move a note to trasharchive_note: Archive a note (removes from main list, keeps searchable)unarchive_note: Unarchive a note
Tag Management
add_tags: Add tags to an existing noteopen_tag: Open Bear and show all notes with a specific tagrename_tag: Rename a tag across all notes
Bear Integration
open_note: Open a specific note in Bearsearch_bear: Open Bear and show search results for a term
All write operations use Bear's official x-callback-url scheme, which requires Bear to be installed and running.
Configuration
You can override the default Bear database path by setting the DB_ROUTE environment variable:
export DB_ROUTE="/path/to/custom/database.sqlite"Development
Running the server directly
cd /Users/borag/Work/mcp-bear
.venv/bin/python -m mcp_bear.serverTesting with MCP Inspector
# Install the MCP inspector
pip install mcp-inspector
# Run the inspector
mcp-inspector python -m mcp_bear.serverArchitecture
mcp_bear/server.py: Main MCP server implementationmcp_bear/database.py: SQLite database access for read operationsmcp_bear/bear_url.py: Bear x-callback-url operations for write operationsUses Python's built-in
sqlite3module (no native dependencies)
Security Note
This server uses parameterized SQL queries to prevent SQL injection attacks, improving upon the original TypeScript implementation.
License
ISC
Author
Bora Gonul me@boragonul.com
Available Tools
16 toolsadd_tagsC
Add tags to an existing note
Inputs: note_id (required), tags (required).
add_textC
Add text to an existing note
Inputs: note_id (required), text (required), mode, open_note.
archive_noteC
Archive a note
Inputs: note_id (required).
create_noteB
Create a new note in Bear
Inputs: title, text, tags, pin, +1 more.
get_archived_notesC
Get all archived notes
Inputs: no named parameters.
get_note_by_idC
Get a specific note by its unique identifier
Inputs: note_id (required).
get_notesB
Get all notes from Bear
Inputs: no named parameters.
get_notes_by_tagB
Get all notes with a specific tag
Inputs: tag (required).
get_notes_likeB
Get notes that include a specific text string
Inputs: like (required).
get_tagsB
Get all note tags. You can search notes by tags with get_notes_like
Inputs: no named parameters.
open_noteC
Open a specific note in Bear
Inputs: note_id (required).
open_tagB
Open Bear and show all notes with a specific tag
Inputs: tag (required).
rename_tagB
Rename a tag across all notes
Inputs: old_tag (required), new_tag (required).
search_bearC
Open Bear and show search results for a term
Inputs: term (required).
trash_noteC
Move a note to trash
Inputs: note_id (required).
unarchive_noteC
Unarchive a note
Inputs: note_id (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 16 tools
Every tool has a clearly distinct purpose with no ambiguity. For example, 'add_tags' and 'add_text' are clearly different operations on notes, while 'get_notes_by_tag' and 'get_notes_like' target different filtering criteria. The descriptions make each tool's specific function immediately apparent.
All tools follow a consistent verb_noun pattern with snake_case throughout (e.g., 'create_note', 'get_notes_by_tag', 'archive_note'). The naming is predictable and readable, with verbs like 'add', 'get', 'open', 'rename', 'search', and 'trash' used appropriately across the set.
With 16 tools, this is well-scoped for a note management server like Bear. Each tool earns its place by covering distinct operations such as CRUD for notes, tag management, archiving, and search, without feeling bloated or insufficient for the domain.
The tool set provides complete coverage for note management in Bear. It includes full CRUD operations (create, read, update via add_tags/add_text, delete via trash_note), lifecycle management (archive/unarchive), tag operations (get, rename), and comprehensive search capabilities (by ID, tag, text, or general search). No obvious gaps exist for typical agent workflows.
Maintenance
Related MCP Connectors
Google Keep-style notes app with an MCP server for AI agents to read/write notes.
Search, read, and write your Apple Notes from ChatGPT/Claude via a local Mac agent + MCP relay.
Markdown-based note-taking with a hosted MCP server. Your notes serve you and your AI.
Hosted MCP server for Obsidian Sync vaults: read, search, edit notes. Not affiliated with Obsidian.
Related MCP Servers
- FlicenseBqualityDmaintenanceA Model Context Protocol server that provides Claude with access to search, retrieve, and analyze notes from the Bear App through natural language queries.78-
- AlicenseAqualityDmaintenanceA Model Context Protocol server that integrates with Bear App, enabling AI assistants to create, search, modify, and organize notes and tags through X-callback-URL scheme.161MIT
- AlicenseAqualityCmaintenanceMCP server for Bear notes via CloudKit. Full CRUD, tags, search, TODOs, attachments, and a context library for curated LLM knowledge bases.3568MIT
- FlicenseNot gradedqualityDmaintenanceAn MCP server that gives AI assistants read and write access to Bear notes on macOS.-