messages-mcp
MCP server for reading and responding to local macOS Messages from clients like Claude Desktop and Codex.
README
messages-mcp
MCP server for reading and responding to local macOS Messages from clients like Claude Desktop and Codex.
The server reads ~/Library/Messages/chat.db, resolves names through local Contacts/AddressBook source databases, and sends through the macOS Messages app using AppleScript.
Requirements
- macOS with Messages configured
- Python 3.10+
- Full Disk Access for the app running the MCP server
- Automation permission for controlling Messages when sending
Install
From this repo:
python3 -m venv .venv
. .venv/bin/activate
pip install -e .
Run manually:
messages-mcp
Claude Desktop
Add this to Claude Desktop's MCP config, adjusting the path:
{
"mcpServers": {
"messages": {
"command": "/Users/YOU/Documents/messages-mcp/.venv/bin/messages-mcp"
}
}
}
Restart Claude Desktop after editing the config.
Codex
Use the command path from the virtualenv:
{
"mcpServers": {
"messages": {
"command": "/Users/YOU/Documents/messages-mcp/.venv/bin/messages-mcp"
}
}
}
Tools
-
list_chats(query="", limit=25)List recent chats, optionally filtered by contact name, phone/email, or group name. -
get_conversation_context(chat, limit=30)Return recent messages and attachment metadata for a chat.chatcan be a contact name, phone/email, or group name. -
get_unread_messages(chat="", limit=20)Return unread incoming messages for one chat or all recent chats. -
send_message(chat, text, dry_run=true)Send a message through Messages.dry_rundefaults totrue; passfalseto actually send. -
mark_read(chat)Best-effort mark a chat as read through Messages.
Safety
Sending is intentionally explicit. The send_message tool defaults to dry_run=true, so clients can inspect the target and text before sending.
Do not run this server for an MCP client you do not trust. Any connected client with tool access can read local Messages context and, if it calls send_message(..., dry_run=false), send messages through your account.
Troubleshooting
If chats do not appear:
- Give the MCP host app Full Disk Access.
- Use a contact name for one-on-one chats.
- Use the exact group chat title for group chats.
- Make sure Messages is signed in and synced locally.
If sending fails:
- Open Messages once manually.
- Approve the Automation permission prompt.
- Try
send_message(..., dry_run=true)first to verify the resolved target.
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.
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.
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.
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.
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.
E2B
Using MCP to run code via e2b.