AgentsMesh/AgentsMesh

The AI Agent Workforce Platform. Run a hundred AI coding agents across your own machines — schedule, isolate, and steer them all from one console.

What it solves

AgentsMesh addresses the operational challenges of scaling AI agent usage from a single user to a large-scale workforce. It solves the "one operator, a hundred agents" problem by providing a control layer that prevents hardware bottlenecks (by distributing agents across multiple machines), eliminates terminal clutter (via a centralized console), prevents state corruption (through isolated workspaces), and stops long-running agents from silently failing (via an autopilot system).

How it works

The platform uses a split architecture separating the control plane from the data plane:

  • Control Plane: A Go-based backend manages authentication, pod lifecycles, and runner certificates using gRPC and mTLS.
  • Data Plane: A stateless Relay cluster handles high-volume terminal I/O via WebSockets, ensuring the backend doesn't bottleneck on PTY traffic.
  • Runner Fleet: Self-hosted daemons installed on various machines spawn "AgentPods"—isolated execution environments containing a PTY terminal and a dedicated Git worktree sandbox.
  • Client Layer: A shared Rust core (compiled to WASM for web/desktop and via UniFFI for iOS) provides a single source of truth for business logic across all interfaces.

Who it’s for

It is designed for developers and team leads who want to run dozens or hundreds of terminal-based AI agents (such as Claude Code, Aider, or Gemini CLI) simultaneously across their own infrastructure without needing to manually manage individual terminals or machines.

Highlights

  • Runner Fleet: Distribute agent workloads across any number of self-hosted machines to scale beyond a single laptop's capacity.
  • Workspace Isolation: Each agent gets its own pod with a private Git worktree and credentials to prevent collision.
  • Autopilot: A control agent that monitors pods and automatically sends the next instruction when an agent goes idle.
  • Mesh & Channels: Allows pods to be bound together in a topology to communicate via @mentions.
  • Multi-platform Console: Unified management via Web, Electron desktop, and iOS apps.

Related

  • Project
    agentscope-ai/AgentTeams

    An open-source collaborative multi-agent runtime platform that uses a Manager-Workers architecture to enable secure, auditable, and human-visible agent collaboration in IM rooms.

  • Project
    builderz-labs/mission-control

    A self-hosted control plane for operating AI agents, allowing users to dispatch tasks, track spend, and coordinate multiple agent runtimes from a single dashboard.

  • Project
    outsourc-e/hermes-workspace

    A comprehensive command center for AI agents that integrates chat, memory, skills, and a terminal into a single workspace for orchestrating multi-agent swarms.

  • Project
    agent-tower/core

    A command center for managing multiple AI agents, featuring a Kanban dashboard, automatic Git Worktree isolation for conflict-free coding, and multi-agent collaboration.

  • Project
    YoanWai/agent-manager

    A terminal-based manager for AI coding agents that allows you to run multiple CLI agents side-by-side with integrated diff review and live status tracking.