iMessage MCP Server
Enables Claude to send and read iMessages on macOS, with smart contact lookup, message history retrieval, and cross-conversation search using natural language commands.
README
iMessage Gateway CLI
The fastest iMessage integration for Claude Code. Direct CLI architecture delivers 19x faster performance than MCP-based alternatives.
Features
- Send Messages: Send iMessages using natural language
- Read Messages: Retrieve message history with contacts or phone numbers
- Smart Contact Lookup: Fuzzy matching for contact names
- Semantic Search (RAG): AI-powered search across iMessages, SuperWhisper, Notes
- Follow-up Detection: Find conversations needing response
- Group Chats: List and read group conversations
- Analytics: Conversation patterns and statistics
Performance
| Operation | Gateway CLI | MCP-based | Speedup |
|---|---|---|---|
| List contacts | 40ms | ~763ms | 19x |
| Find messages | 43ms | ~763ms | 18x |
| Unread messages | 44ms | ~763ms | 17x |
| Groups | 61ms | ~763ms | 12x |
| Semantic search | 150ms | ~900ms | 6x |
Requirements
- macOS (required - iMessage is macOS only)
- Python 3.9+
- Full Disk Access permission (for reading message history)
Quick Start
1. Clone and Install
git clone https://github.com/yourusername/imessage-gateway.git
cd imessage-gateway
pip install -r requirements.txt
2. Set Up Contacts
# Sync from macOS Contacts (recommended)
python3 scripts/sync_contacts.py
# Or manual setup
cp config/contacts.example.json config/contacts.json
# Edit config/contacts.json with your contacts
3. Grant Permissions
-
Full Disk Access (for reading messages):
- System Settings → Privacy & Security → Full Disk Access
- Add Terminal.app or your Python interpreter
-
Automation (for sending messages):
- Will be requested automatically on first send
4. Test It Out
# List your contacts
python3 gateway/imessage_client.py contacts --json
# Check unread messages
python3 gateway/imessage_client.py unread --json
# Get recent messages
python3 gateway/imessage_client.py recent --limit 20 --json
# Send a message
python3 gateway/imessage_client.py send "John" "Hey, are you free for coffee?"
Command Reference (27 Commands)
Messaging (3)
# Send to contact
python3 gateway/imessage_client.py send "John" "Hello!"
# Send to phone number directly
python3 gateway/imessage_client.py send-by-phone "+14155551234" "Hi there!"
# Add a new contact
python3 gateway/imessage_client.py add-contact "Jane Doe" "+14155559876"
Reading (12)
# Messages with a contact
python3 gateway/imessage_client.py messages "John" --limit 20 --json
# Find messages (keyword search)
python3 gateway/imessage_client.py find "John" --query "meeting" --json
# Recent across all contacts
python3 gateway/imessage_client.py recent --limit 50 --json
# Unread messages
python3 gateway/imessage_client.py unread --json
# Recent phone handles
python3 gateway/imessage_client.py handles --days 30 --json
# Messages from unknown senders
python3 gateway/imessage_client.py unknown --days 7 --json
# Attachments
python3 gateway/imessage_client.py attachments "John" --json
# Voice messages
python3 gateway/imessage_client.py voice --json
# Links shared
python3 gateway/imessage_client.py links --days 30 --json
# Message thread
python3 gateway/imessage_client.py thread "<message-guid>" --json
# Scheduled messages
python3 gateway/imessage_client.py scheduled --json
# Conversation summary
python3 gateway/imessage_client.py summary "John" --days 7 --json
Groups (2)
# List group chats
python3 gateway/imessage_client.py groups --json
# Messages from a group
python3 gateway/imessage_client.py group-messages --group-id "chat123456" --json
Analytics (3)
# Conversation analytics
python3 gateway/imessage_client.py analytics "John" --days 30 --json
# Follow-ups needed
python3 gateway/imessage_client.py followup --days 7 --json
# Reactions
python3 gateway/imessage_client.py reactions "John" --json
Contacts (1)
# List all contacts
python3 gateway/imessage_client.py contacts --json
Semantic Search / RAG (6)
# Index iMessages for semantic search
python3 gateway/imessage_client.py index --source=imessage --days 30
# Index SuperWhisper transcriptions
python3 gateway/imessage_client.py index --source=superwhisper
# Index Notes
python3 gateway/imessage_client.py index --source=notes
# Index all local sources
python3 gateway/imessage_client.py index --source=local
# Semantic search
python3 gateway/imessage_client.py search "dinner plans with Sarah" --json
# AI-formatted context
python3 gateway/imessage_client.py ask "What restaurant did Sarah recommend?"
# Knowledge base stats
python3 gateway/imessage_client.py stats --json
# List available/indexed sources
python3 gateway/imessage_client.py sources --json
# Clear indexed data
python3 gateway/imessage_client.py clear --source=imessage --force
Architecture
Messages.db (SQLite) ←─ ~/Library/Messages/chat.db
↓
MessagesInterface ────→ SQLite queries (read)
│ AppleScript → Messages.app (send)
↓
ContactsManager ──────→ config/contacts.json (fuzzy matching)
↓
Gateway CLI ──────────→ gateway/imessage_client.py
↓
Claude Code ──────────→ Bash tool calls
Why Gateway CLI?
The Gateway CLI architecture bypasses MCP framework overhead entirely:
- Direct execution: Python script runs immediately, no JSON-RPC initialization
- No session startup: MCP servers have ~700-800ms cold start per session
- Smaller footprint: No MCP SDK dependency, simpler codebase
- Same reliability: Uses identical
MessagesInterfacecode for all operations
Claude Code Integration
Using the Skill
The imessage-gateway skill provides natural language access:
/imessage-gateway unread
/imessage-gateway send John "Running late!"
/imessage-gateway search "meeting next week"
Bash Pre-approval
Ensure your settings include:
Bash(python3:*::*)
Configuration
Contact Format
Contacts are stored in config/contacts.json:
{
"contacts": [
{
"name": "John Doe",
"phone": "14155551234",
"relationship_type": "friend",
"notes": "Optional notes"
}
]
}
Phone numbers can be in any format - they're normalized automatically.
Troubleshooting
"Contact not found"
- Run
python3 scripts/sync_contacts.pyto sync contacts - Check
config/contacts.jsonexists and has contacts - Try partial names (e.g., "John" instead of "John Doe")
"Permission denied" reading messages
- Grant Full Disk Access to Terminal/Python
- Restart Terminal after granting permission
- Verify:
ls ~/Library/Messages/chat.db
Messages show "[message content not available]"
- Some older messages use a different format
- Attachment-only messages don't have text content
- This is normal for some message types
Development
# Run tests
pytest tests/ -v
# Run performance benchmarks
python3 -m Texting.benchmarks.run_benchmarks
# Sync contacts
python3 scripts/sync_contacts.py
Privacy & Security
- All data stays local on your Mac
- No cloud services for core functionality
- Contacts file is gitignored by default
- Message history accessed read-only
- Optional OpenAI API for semantic search embeddings
MCP Server (Archived)
The MCP server has been archived in mcp_server_archive/. See mcp_server_archive/ARCHIVED.md for restoration instructions if needed.
License
MIT License - see LICENSE for details.
Built for use with Claude Code
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.
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.
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
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.
E2B
Using MCP to run code via e2b.