mcp_agent_mail
Asynchronous coordination layer for AI coding agents: identities, inboxes, searchable threads, and advisory file leases over FastMCP + Git + SQLite
What is mcp_agent_mail?
Asynchronous coordination layer for AI coding agents: identities, inboxes, searchable threads, and advisory file leases over FastMCP + Git + SQLite
Overview of mcp_agent_mail
What it is
MCP Agent Mail is a mail-like coordination layer for AI coding agents, exposed as an HTTP-only FastMCP server. It provides agents with memorable identities, an inbox/outbox, searchable message history, and voluntary file reservation "leases" to avoid conflicts. The system is backed by Git for human-auditable artifacts and SQLite for indexing and queries. This Python implementation is superseded by the Rust rewrite (mcp_agent_mail_rust) and is no longer actively maintained; new installs should use the Rust version.
Key features
- Register a temporary-but-persistent identity (e.g., GreenCastle)
- Send and receive GitHub-Flavored Markdown messages with images
- Search, summarize, and thread conversations
- Declare advisory file reservations (leases) on files or globs to signal intent
- Inspect a directory of active agents, programs/models, and activity
- Backed by Git (auditable artifacts) and SQLite (indexing and queries)
Use cases
Modern projects often run multiple coding agents at once (backend, frontend, scripts, infra). Without a shared coordination fabric, agents overwrite each other's edits, miss critical context from parallel workstreams, or require humans to liaison messages across tools. MCP Agent Mail provides a lightweight, interoperable layer so agents can coordinate across one or more codebases using FastMCP clients and CLI tools (Claude Code, Codex, Gemini CLI, Factory Droid, etc.).
Getting started
One-line installer:
curl -fsSL "https://raw.githubusercontent.com/Dicklesworthstone/mcp_agent_mail/main/scripts/install.sh?$(date +%s)" | bash -s -- --yes
After installation, start the server from anywhere by typing:
am
For manual installation, clone the repository, set up a Python 3.14 virtual environment with uv, and run the server script.
License and pricing
This project is open source. The license is not stated in the metadata (NOASSERTION).
Based on the project's public README and repository metadata.