OmniFocus MCP Server
Enables comprehensive management of OmniFocus on macOS through 17 specialized tools for projects, tasks, and organization. Users can create, update, and filter items or navigate the interface using natural language via the Model Context Protocol.
README
OmniFocus MCP Server
An MCP (Model Context Protocol) server that provides tools for interacting with OmniFocus on macOS.
Note: This is a personal project. You're welcome to fork, clone, and adapt it for your own use, but I'm not accepting pull requests or feature requests. Issues are used for internal tracking.
Features
This server provides 21 comprehensive tools for managing OmniFocus (v0.8.3 API):
Project Management (5 tools)
- get_projects - Get all projects with filtering (by ID, query, status) and optional full notes
- create_project - Create new projects with optional folder placement and review intervals
- update_project - Update single project (all properties: name, note, status, folder, review interval, etc.)
- update_projects - Batch update multiple projects (safe fields only: status, folder, review settings)
- delete_projects - Delete one or multiple projects (accepts single ID or list)
Task Management (6 tools)
- get_tasks - Get tasks with comprehensive filtering (by ID, project, parent, tags, status, dates, flags, text search, inbox-only)
- create_task - Create tasks with all properties (due dates, defer dates, flags, tags, estimated time, notes, parent tasks)
- update_task - Update single task (all properties: name, note, dates, flags, tags, status, project, parent, etc.)
- update_tasks - Batch update multiple tasks (safe fields only: dates, flags, tags, status, project)
- delete_tasks - Delete one or multiple tasks (accepts single ID or list)
- reorder_task - Change task order relative to siblings (before/after positioning)
Folder & Organization (2 tools)
- get_folders - Get all folders with hierarchy
- create_folder - Create folders with optional parent folder
Tags (4 tools)
- get_tags - Get all available tags
- create_tag - Create new tags with optional parent nesting
- update_tag - Update tag name or active status
- delete_tags - Delete one or multiple tags
Perspectives (2 tools - OmniFocus Pro)
- get_perspectives - List all perspectives with type and ID information
- switch_perspective - Switch to a different perspective view
UI Navigation (2 tools)
- set_focus - Focus on one or more projects/folders, or clear focus
- get_focus - Get the currently focused items
Key Changes in v0.8.3:
- v0.8.3: Enhanced focus (multi-item, get_focus), enriched perspectives, tag CRUD, blind agent eval (36/36), docstring improvements
- v0.8.2: AppleScript injection hardening, consistent server error handling, dead code removal, complexity refactoring
- v0.8.1: 35x get_projects() performance optimization via batch AppleScript reads, tag filtering fix, integration test infrastructure improvements
- v0.7.3: Performance benchmarks, release workflow automation, scripts audit, project review optimizations (task health, opt-in lastActivityDate, AppleScript query filter)
- v0.7.2: CHANGELOG date validation, TaskPaper import for fixtures, performance optimization for get_tasks/get_projects, test fixture refactoring complete, test count synchronization
- v0.7.1: Test fixtures infrastructure with automatic cleanup, E2E test refactoring, ~7.8× performance improvement
- v0.7.0: Added set_focus() UI navigation tool, fixed AppleScript focus/tab management bugs
- v0.6.7: Fixed unit test timeout issue, fixed test coverage check script
- v0.6.6: Release process infrastructure, interactive quality check slash commands, workflow enforcement and documentation
- v0.6.5: Fixed AppleScript DONE status bug, added GitHub Actions CI, comprehensive testing documentation
- v0.6.3-v0.6.4: Trunk-based workflow with RC tags, comprehensive hygiene checks for releases
- v0.6.2: Added Claude Code hooks for automated workflow enforcement (prevents commits to main, monitors CI failures)
- v0.6.0: Consolidated 40+ functions into 16 comprehensive tools. All updates now go through
update_task()andupdate_project()instead of specialized functions. See CHANGELOG.md for migration guide.
Prerequisites
- macOS (OmniFocus is macOS-only)
- OmniFocus app installed
- Python 3.10 or higher
- UV package manager (recommended) or pip
Installation
Stable Release (Recommended)
Install from a specific version tag for stability:
# Clone this repository
git clone https://github.com/s-morgan-jeffries/omnifocus-mcp.git
cd omnifocus-mcp
# Checkout latest stable release
git checkout v0.8.3 # Or latest version from releases
# Option 1: Using UV (recommended)
curl -LsSf https://astral.sh/uv/install.sh | sh
uv pip install -e .
# Option 2: Using pip
python3 -m venv venv
source venv/bin/activate
pip install -e ".[dev]"
Development / Latest Features
Use the main branch for the latest unreleased features (always releasable, tests pass):
# Clone this repository
git clone https://github.com/s-morgan-jeffries/omnifocus-mcp.git
cd omnifocus-mcp
# main branch is default - always releasable but may have unreleased features
# Install with UV or pip (same as above)
Note: Main branch follows trunk-based development - it's always releasable and tests pass, but may contain features not yet included in a tagged release. For production use, install from a specific version tag.
Configuration
For Claude Desktop
Add this server to your Claude Desktop configuration file:
Location: ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"omnifocus": {
"command": "uv",
"args": [
"--directory",
"/absolute/path/to/omnifocus-mcp",
"run",
"omnifocus-mcp"
]
}
}
}
Or if using a virtual environment directly:
{
"mcpServers": {
"omnifocus": {
"command": "/absolute/path/to/omnifocus-mcp/venv/bin/python",
"args": [
"-m",
"omnifocus_mcp.server_fastmcp"
]
}
}
}
Important: Replace /absolute/path/to/omnifocus-mcp with the actual absolute path to this directory.
For Other MCP Clients
Any MCP-compatible client can use this server by launching it with:
python -m omnifocus_mcp.server_fastmcp
The server communicates via stdin/stdout using the MCP protocol.
Usage Examples
Once configured, you can ask Claude (or any MCP client) to interact with OmniFocus:
Project Management:
- "Show me all my OmniFocus projects"
- "Search for projects related to 'budget'" → uses
get_projects(query="budget") - "Create a new project called 'Q1 Planning' in my Work folder"
- "What projects are due for review?"
Task Management:
- "Add a task 'Review Q4 numbers' with a due date next Friday, flagged as important"
- "Show me all available tasks that are overdue"
- "Do I have a mortgage payment due this week?" → uses
get_tasks(query="mortgage", due_relative="this_week") - "Mark task task-001 as complete"
- "Move this task to my Personal project"
- "Set time estimate for task-001 to 90 minutes"
Inbox & Quick Capture:
- "Add 'Call dentist' to my inbox"
- "Show me what's in my inbox" → uses
get_tasks(inbox_only=True)
Organization:
- "Add the 'urgent' tag to task-001"
- "Create a new folder called 'Personal Goals' inside 'Life'"
- "Make task-002 a subtask of task-001"
GTD Review:
- "Set the review interval for this project to 2 weeks"
- "Mark project proj-001 as reviewed"
Perspectives (Pro):
- "Switch to my 'Daily Worklist' perspective"
- "What perspectives do I have available?"
Development
Running Tests
The project has comprehensive test coverage with 544 tests (406 passing, 138 skipped). See docs/guides/TESTING.md for detailed breakdown and procedures.
# Activate virtual environment
source venv/bin/activate
# Run all tests
pytest tests/
# Run with coverage report
pytest tests/ --cov=src/omnifocus_mcp --cov-report=term-missing
# Run specific test file
pytest tests/test_omnifocus_client.py -v
Testing the Server
You can test the server locally:
# Run the server (it will wait for MCP protocol messages on stdin)
python -m omnifocus_mcp.server_fastmcp
Debugging
The server logs to stderr, which Claude Desktop captures. To see logs:
- Open Claude Desktop
- Go to the menu and enable developer mode
- View logs in the developer console
Permissions
The first time you run this server, macOS may prompt you to grant permissions for:
- Accessibility access (to control OmniFocus via AppleScript)
- Automation permissions for OmniFocus
Grant these permissions to allow the server to function properly.
Troubleshooting
"Operation not permitted" errors
- Check System Settings > Privacy & Security > Automation
- Ensure the terminal or Claude Desktop has permission to control OmniFocus
"OmniFocus is not running" errors
- Make sure OmniFocus app is running
- Try restarting OmniFocus and the MCP server
Server not appearing in Claude Desktop
- Check the configuration file syntax (must be valid JSON)
- Verify the absolute paths are correct
- Restart Claude Desktop after changing configuration
- Check Claude Desktop logs for error messages
License
MIT
Contributing
Contributions welcome! Please see CONTRIBUTING.md for:
- Development workflow and TDD requirements
- Before-commit checklist
- Code quality standards
- How to run tests
For quick contributions: Open an issue or PR with your proposal.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.