Agent frameworks & infrastructure
Multi-agent orchestration
6,185 agents · most starred first
Multi-agent orchestration, page 22 of 62
Agent Delegated Authority Protocol (ADAP) v0.1 Public Draft. An open protocol for delegated authority between AI agents.
🦀 A self-hosted, multi-agent service framework built in Rust. Agents as services, not scripts.
Multi-agent orchestration system for GitHub Copilot with specialized agents, reusable skills/prompts, and validation tooling.
Production-ready AI agent configurations and orchestration patterns
Repo for HAICOSYSTEM
Experimenting with meta-learning approaches to opponent modelling in MARL. Building upon previous public implementations of MADDPG and M3DDPG.
Desktop control center for OpenClaw — Multi-instance orchestration, team collaboration, real-time chat & observability (Tauri + React)
Remote Agent Invocation Layer
T20: Multi-Agent. Orchestrator-delegate model. TAS. Goal -> plan -> delegate. Agents: Gemini family. Autonomous, traceable. Logs sessions. CLI. Usage: `t20-cli "goal string"`. Artifacts of high value.
Boilerplate for multi-LLM agent applications using Langchain v3 with multi-model reasoning capabilities.
Production-ready FastAPI service for concurrent Claude Agent session management with real-time monitoring, cost tracking, and session resumption via HTTP API.
First-class protocol and framework for enabling collaborative intelligence and communication across specialized Data Agents.
Agent Message Transfer Protocol is a federated communication protocol designed for reliable agent-to-agent communication across organizational boundaries.
Coordinate your AI agent team. Shared tasks, memory, reflections, and presence. Self-host for free.
CARD: Towards Conditional Design of Multi-agent Topological Structures. Published at ICLR 2026.
A powerful framework for deploying AI agents with multi-agent support, session management, event streaming, and LLM-agnostic integration.
🚀 Build AI Agent Teams as Production-Ready APIs. Orchestrate CrewAI agents with FastAPI for enterprise-grade AI services. Leverage Groq's lightning-fast LLMs to deploy collaborative AI workflows at scale.
Multi-agent Claude Code you can actually control. Spawn specialist teams in tmux, watch them work side-by-side, steer when needed. Built on native Claude Code agents. No black boxes.
Multi-LLM AI debate platform — orchestrate structured debates between Claude, GPT, and Gemini with consensus scoring and synthesized verdicts.
Cord — distributed agent fabric for LLMs, MCP servers and AI agents. Discover any LLM CLI / HTTP backend / MCP server across machines via natural-language semantic search. Built in Rust.
OpenClaw Multi Agent / 一人公司 / AI 版聊天IM
Multi-model AI agent orchestrator. Run Claude, GPT, and Ollama agents in one workspace with group chat, smart routing, and per-agent model assignment.
This project demonstrates Agent-to-Agent (A2A) conversations using LangGraph with the A2A protocol. It includes implementations in both **Python** and **TypeSc…
Sema is a content-addressed vocabulary for AI agents. Precise thinking, reliable references, safe multi-agent coordination. Hash the meaning, get the word.
Multi-agent coordination system for software engineering with 9 specialized AI agents - Documentation, API Design, Backend Architecture, Java/Go Development, DDD, and Observability
Personal AI agent orchestration runtime for Claude Code: build, schedule, and run private TypeScript agents with A2A, launchd, and an Electron UI.
Open-source agent orchestration framework providing deterministic multi-agent workflows, AST verification gates, and maintainer automation pipelines.
🤖 Self-Evolving Multi-Agent AI System | 116+ Skills | TrendRadar · Reviewer | Learn, Build & Evolve
Kimi K2.7 Swarm Workstation — 300-agent parallel AI hive. Beats Claude Opus 4.8 on tool use (81.1% MCPMark), 5× cheaper. Win/Mac, one-click.
IM-native multi-agent collaboration platform — @mention Claude Code, Codex, OpenCode like teammates
Pi extension for parallel panel review and judge synthesis
P2P multi agent orchestration harness, universal automation system and IDE
Serve any ACP coding agent — Claude Code, Codex, opencode — over the A2A protocol, with tool progress, file diffs, and permission prompts preserved.
Local-first control plane for long-running and multi-agent AI systems. Observe, steer, recover and orchestrate agents.
Assemble the AI council you can't afford to hire. Deliberon — structured multi-agent AI deliberation for Windows. Installer downloads + release notes.
Open communication and coordination for AI agents.
K2 by Alakazam Labs: AI agents, chat, mail, documents and apps on a server you own. Free and MIT licensed.
A long-horizon agent orchestration toolkit in OCaml implementing Cordis-like extensibility and graceful fallbacks
Skills to teach agents from different vendors call each other from CLI. Let Claude prompt Codex and Pi as Subagent.
Live agent-to-agent communication for Pi — local broker, mesh.v1 protocol, honest delivered/read/answered statuses, hash-only ledger. Zero runtime deps. Install: pi install npm:pi-mesh
Community Codex Multi-Agent V2 arrangements: lead agents, specialist TOMLs, orchestration skills and optional hooks.
A coordination protocol for small agent swarms, and the pi extension that runs it.
Several coding agents, one repo, at the same time: a reproducible lab to measure how they coordinate, and Médula, a kernel that decides who waits for whom.
Agent workbench for auto research, self-improvement loops, and long-running work — multi-agent teams, persistent project memory, cross-machine GPU training. Developed entirely inside Cortex.
Your Paperclip agent company as a live pixel office: every agent at a desk, work moving between them, and a Decision box for everything waiting on you.
An interactive command-line interface that uses the Agent2Agent (A2A) & Model Context Protocols (MCP) to manage multi-agent systems.
BullX is an AgentOS for working side by side with self-directed AI Colleagues.
Multi-agent AI assistant platform with specialized agents for coding, finance, news, real estate, travel, image generation, and shopping.
Box Jump is a co-operative multi-agent reinforcement learning environment!
Oracle Generative AI Multi Agent Reasoning Orchestrator
MSc Informatics dissertation project - University of Edinburgh: Curiosity in Multi-Agent Reinforcement Learning
A durable broker and replayable AI workflow runtime with guardrails, governance, and multi-agent orchestration.
Openclaw 数字人孵化平台 · 飞书多智能体管理器
File-based coordination protocol for stateless AI models
Collaborative Multi-Agent AI System with LangGraph
Give your AI agents a chief of staff.
LangGraph Multi-Agent Handoffs, Clearly Explained
The enterprise-grade CrewAI alternative for Go. Orchestrate strictly-typed, parallel AI agents with LangGraph-style Flow persistence, distributed A2A protocols, Model Context Protocol (MCP), and Human-in-the-Loop training.
A fully local multi-agent AI board — CEO, CTO, CFO, CPO, COO running on your own GPU/CPU. No cloud. Just Local Models.
AI Company Orchestration Platform — Deploy governed teams of autonomous AI agents with Runtime Agentic IAM, credential vaults, heartbeat orchestration, and enterprise security. Built on Paperclip.
Organize your AI agent teams with the Viable System Model. The missing organizational layer for multi-agent systems.
13 specialized AI agents — PM, Architect, DBA, UI Designer, Project Manager, Frontend, .NET, QA, DevOps, Planner, Java, Python, DigitalTeam(coordinator) — forming a structured software delivery team with defined handoffs, gate reviews, and workflow. supports Github Copilot, Claude Code, OpenAI Codex CLI and Trae.
Quorum: orchestrate a swarm of AI experts on any question. Specialists debate, research, and validate — then a polymath supervisor delivers the verdict. One command, multiple minds, stress-tested answers.
SlackAgents is a multi-agent library for building AI agents to automate daily work in Slack workspaces
Connect agents across runtimes, frameworks, and teams.
Raven is a self-hostable team Agent platform providing isolated workspaces and unified runtimes for multi-agent workflows.
UNS-native, cybersecurity-aware multi-agent reference architecture for U.S. small-to-mid manufacturers. Industrial agentic intelligence aligned with ISA-95, NIST AI RMF, and CMMC L2.
Elixir framework for multi-model LLM councils. Debate, vote, tournament, synthesize. Provider-agnostic.
A runtime-portable 5-agent adjudicator council for universal quality gating of text artifacts.
An AI-powered workforce platform that executes multi-agent tasks with automatic provider failover and CEO agent orchestration. It detects technology stacks across 1,000+ platforms and provisions industry-specific specialists while running on customer hardware with no data leaving their network.
streamlit region:us
gradio region:us
AgentWorks: the open-source control plane for running, measuring, and improving AI agent workflows across your company.
A minimalist, discrete multi-agent predator-prey archytype environment.
✨ Intelligent, adaptive multi-agent orchestration & deterministic zero-cost telemetry engine for Codex with native macOS floating widget
Stand up your own cross-model AI engineering team — the way Neo.mjs runs its own. Named AI maintainers with persistent identity, shared memory and review across labs, coordinating as a team of teams on your codebases. Multi-window, in the browser or an Electron shell.
Reusable Agent Team presets for DeepSeek Harness: a captain plus members with their own prompt, model, reasoning effort, and tools.
Run a team of AI coding agents, such as Claude Code and Codex, on one project. They hand work to each other and review it across AI companies; a supervisor can keep them running; and a shared, curated memory brings lessons from past work back into later turns. Their messages and records are files in your project: no server, no account.
Make your best coding agent the CTO: it plans, reviews and merges while Codex, Grok, Claude Code, Copilot & co. do the work in parallel git worktrees. An Agent Skill + a few hundred lines of bash.
Enterprise Capability Planning Compiler for governed, runtime-neutral business plans.
Harmovela Protocol: open coordination protocol for autonomous systems — event, task, state, delegation, recovery, and governance.
CARD: Towards Conditional Design of Multi-agent Topological Structures. Published at ICLR 2026.
Multi-agent governance layer for self-maintaining LLM wikis
The Universal Agent Interoperability Protocol: The TCP/IP of the AI Economy
Hive is a Multi Agent Orchestrator modelled after Agile Teams
Mandali (मंडली) — Autonomous multi-agent adversarial collaborative system using GitHub Copilot SDK
A framework to orchestrate multiple AI agents, enabling them to collaborate efficiently through smart threading, dependency management, context sharing, and distributed execution via a built-in swarm system. Built on top of smolagents, it allows the creation of an AI dream team that can scale to tackle complex tasks.
Universal Agent Protocol (UAP) - Making AI agents work together seamlessly. The standard way for agents to find, talk to, and coordinate with each other in multi agent systems (MAS).
Multi-agents operating system layer on top of Claude Code
Containerized A2A-native agents with simple setup
A multi-agent system built with DSPy that coordinates research and writing agents to produce comprehensive articles on any topic.
Seamlessly integrate IoT data with AI agents, enabling the effortless parsing, processing, and utilization of IoT data streams.
Multi-agent orchestration on top of Claude Code. Supervisor delegates, specialists execute, you gate. Browser-based IDE; local-first.
A Multi-Agent Research System using Claude Code Skills. it is inspired by Anthropic claude-agent-sdk-demos
A collection of demonstrations showcasing different patterns for implementing multi-agent workflows using LangGraph. Each example highlights specific orchestration approaches to help developers understand and build collaborative AI systems.
Omnimesh - Universal network of agents and services
Build and deploy a multi-agent AI system on AWS using Bedrock AgentCore, 4 agents collaborate to research, analyze, and write reports
Multi-mode, multi-agent idea ➡️ reality creation app.
AWS Bedrock Agentcore sample illustrating multi tenancy