Repolode
Submit
23blocks-OS

23blocks-OS/ai-maestro

A self-hosted dashboard to orchestrate Claude Code, Codex and other AI agents, with persistent memory, agent-to-agent messaging and multi-machine support.

View on GitHubWebsite
Stars
833
Forks
116
Open issues
23
Last commit
22 hours ago

What is it?

AI Maestro is a self-hosted dashboard for running many AI agents from one place. Agents can be Claude Code, Codex, Grok Build, or any terminal-based AI, and they appear with their status, role and team, whether they run in local tmux sessions, Docker containers, or on AWS EC2 and ECS Fargate.

Each agent keeps a long-term memory built from its past conversations, and agents communicate directly through the Agent Messaging Protocol (AMP). Machines join a peer mesh without a central server, so agents on a Mac and a Linux box can be seen and coordinated together. The project also includes a Kanban board, a team meeting view, chat with approval cards, and a secrets vault that lets agents use credentials without seeing their values.

AI Maestro is MIT licensed and bundles a Claude Code plugin with 14 skills for memory search, code graph queries, task boards and messaging.

Key features

  • 01Dashboard showing Claude Code, Codex and other agents across local, Docker and AWS hosts
  • 02Long-term memory built from each agent's past conversations, with an entity graph
  • 03Agent-to-agent messaging over AMP, with signatures and push notifications
  • 04Peer mesh across machines with no central server
  • 05Secrets vault so agents can use credentials without seeing the values
  • 06Kanban board, team meetings and per-agent task tracking

Use cases

  • Coordinating a team of 3 or more coding agents that currently need manual copy-paste between terminals
  • Running backend, frontend and infrastructure agents on different machines under one dashboard
  • Connecting agents to Slack, Discord or email through a gateway without exposing your network
  • Scheduling recurring agent jobs such as nightly dependency checks

Quick start

  1. 1

    Run the one-line install script from the README

  2. 2

    Open the dashboard in your browser

Things to know

  • Requires Node.js 18+, tmux, and 4 GB of RAM to build.
  • Windows is supported only through WSL2.
  • Install uses a curl-piped-to-bash script, so review it first if that concerns you.
  • Long-term memory and some features depend on Claude Code and optional plugins.

In collections

Claude Code tools, skills & pluginsOpen-source AI agent frameworks & appsAI agent skill librariesSelf-hosted apps

Topics

#ai-agents#multi-agent-systems#agent-orchestration#claude-code#agent-messaging#ai-memory#self-hosted#tmux#agent-communication#agent-skills

FAQ

What is ai-maestro?
A self-hosted dashboard to orchestrate Claude Code, Codex and other AI agents, with persistent memory, agent-to-agent messaging and multi-machine support. AI Maestro is a self-hosted dashboard for running many AI agents from one place. Agents can be Claude Code, Codex, Grok Build, or any terminal-based AI, and they appear with their status, role and team, whether they run in local tmux sessions, Docker containers, or on AWS EC2 and ECS Fargate.
Is ai-maestro free to use? What license does it have?
ai-maestro is open source under the MIT license. Check the license terms before using it in commercial products.
Is ai-maestro still maintained?
Actively maintained. The last commit to ai-maestro was on Oct 21, 2026.
What can I use ai-maestro for?
Coordinating a team of 3 or more coding agents that currently need manual copy-paste between terminals Running backend, frontend and infrastructure agents on different machines under one dashboard Connecting agents to Slack, Discord or email through a gateway without exposing your network Scheduling recurring agent jobs such as nightly dependency checks
How do I install ai-maestro?
Run the one-line install script from the README Open the dashboard in your browser
Where can I get ai-maestro?
The source code and installation instructions are on GitHub: https://github.com/23blocks-OS/ai-maestro

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects