multi-agent

Open-source shared memory layer for multi-agent systems, built around signed, cite-able facts that can be read without an account and integrated via MCP, Python, TypeScript, or REST.

An open-source project that runs full games of Diplomacy between 7 LLM-controlled powers, with negotiation rounds, validated orders, and saved transcripts for replay and analysis.

Open-source voice-driven multi-agent assistant for building real projects on a Mac, with voice, web chat, and WhatsApp channels plus CLI-based task runners.

SwarmWright is a self-hosted multi-agent AI orchestration platform for designing, running, and inspecting swarms of agents from markdown and config. The page highlights Docker-based setup, topology-enforced agent graphs, human approval steps, local storage, and detailed audit logs.

Open-source multi-agent harness from Microsoft for running large self-organized agent teams asynchronously. The repo describes scaling to 1,024 agents, includes a reproduction guide, Docker Compose setup, launch scripts, and an evaluation write-up.

Open-source Python framework for building multi-agent LLM apps that can run on local models through Ollama or on hosted APIs like OpenAI, Anthropic, and OpenRouter. The repo highlights built-in plugins, workflow orchestration, agent serving over HTTP, and guardrails for tool use.

Openvurp is a local-first multi-agent workspace where agents have jobs, tools, memory, and can consult each other instead of answering in isolation. The repo shows browser, CLI, and Docker setup, plus concrete workflows for running commands, writing files, and using multiple channels.

Crew is a local-first collaborative IDE for people and AI agents to work together on projects. It supports multiple agents at once, shared threads, project folders, token tracking, and built-in browser tools like terminal, files, preview, web, and review.

Abralo is a native desktop app for running multiple Claude Code agents at once in one window. It focuses on keeping several agents visible, showing token usage, and surfacing when an agent needs attention.

A cross-platform agent skills repository for consensus-based R&D and repo automation. It describes a multi-agent loop that generates independent plans, converges through a meta-judge, and supports platforms like Claude Code, Codex, Cursor, and Gemini.

Open-source Go runtime for autonomous multi-agent systems. Swarm models work as durable state machines, persists state, supports replay/fork, and integrates with multiple LLM backends.

Open-source Claude Code harness that replaces context compaction with a verified capsule, /clear, and resume flow. It targets long-running and multi-agent sessions with persistent memory stored in git, plus a portable plugin mode and install scripts.

A research project exploring multi-agent LLM coordination in Pokémon Showdown. It compares talk modes, reports win rates and token costs, and frames the environment as a benchmark for studying collaboration between agents.

Kodama is a portable, provider-neutral multi-agent pack for the Kiro CLI. It ships one orchestrator plus eight specialist agents, with install options for Kiro Crew or standalone use, and includes CLI commands for memory, health checks, updates, and uninstall.

Open-source platform for running multiple AI agents alongside teams across chat and dev workflows, with self-hosted deployment and role-based agent coordination.

Custodian Labs is a developer platform for building and deploying AI agents from a small code snippet. The page highlights agent deployment, built-in RAG, model flexibility, multi-agent routing, and a privacy-focused Guardian layer for handling PII.

A plugin for coding agents that lets them message and coordinate with other agents across sessions, including Claude Code, Codex, and Antigravity.

A Rust CLI and MCP server for orchestrating multiple coding agents, fusing their answers, and resuming or routing agent sessions locally.

A Claude Code skill for multi-agent code review that produces findings.md with severity, impact category, verification artifacts, and fix directions.

A file-based protocol and CLI for coordinating multiple AI agents across runtimes, with typed messages, review loops, and shared memory on the local filesystem.

Open-sourced local multi-agent OS that runs three agents on consumer hardware, lets them form goals from environmental pressure, and supports self-modifying workflows through tool creation and change requests.

A session manager for AI coding agents that lets developers run and monitor multiple agents in parallel from a TUI or web dashboard. It supports several agent frameworks, uses git worktrees and optional Docker sandboxes, and includes remote phone access plus PWA-style web use.

Herd is a lightweight desktop IDE for coordinating multiple AI coding agents at once. It supports parallel tasks, real-time status monitoring, and works with tools like Claude Code and Codex on Windows, macOS, and Linux.

A Java framework for building stateful multi-agent workflows on top of Spring AI, with graphs, routing, retries, checkpoints, and shared state.