Agent frameworks & infrastructure
Multi-agent orchestration
6,185 agents · most starred first
Multi-agent orchestration, page 20 of 62
Leave agents running without giving up control.
Beastmode: MofA (Mixture of Agents) orchestration framework for Hermes/OpenClaw/Codex with MemroOS-style context continuity.
Delegate tasks from chat to an isolated workspace, then selectively return results with the next user turn.
An agent orchestrator where your agents are weird alien dogs.
Graph-native studio, runtime integration, and self-hosted observability for mutable multi-agent LLM workflows.
Persisted multi-agent workflow orchestration — DAG tasks, git worktrees, merge gates, desktop + CLI + Slack
Chinese template for designing Multica multi-agent + skill workflows around task type, budget, models, and verification gates.
Spec-driven multi-agent teamwork for any skills-capable agent (Claude Code, Codex, and more) — plan, dispatch, monitor, QA. No framework lock-in.
The open interoperability layer for AI agents. Connect your favorite agents / harnesses and robots. Work together.
Open-source local multi-agent platform
Hive mind mechanics for agents. A step closer towards AGI
Operating Swarm (OS): provider-agnostic agent operating layer and harness. Native execution, external harnesses, and peer OS instances as one abstraction.
Coordinate coding agents in Herdr with assign, handoff, result collection, and follow-up messages. Install with npx skills.
A virtual company of AI employees that runs on your own machine.
Watch a request turn into finished work. AI agents produce and check it
The Company-as-a-Service platform. Build, ship, and scale -- powered by AI teams.
MOCO - Multi-agent Orchestration COre: Lightweight AI agent framework for Python
Multi-agent reinforcement learning on trains, for Deep Learning class at UNIBO
A reusable multi-agent framework built on top of the Cloudflare Agents SDK.
Universal Agent Messaging — encrypted agent-to-agent communication protocol
Docker-based multi-agent AI team. One config file. Unlimited agents. Always on.
Exploring and demonstrating OpenAI's Swarm framework
The Expert Orchestrator AI: Dynamically Adapting, Budget-Aware, and Precisely Tailored to Your Needs
Next-generation multi-agent AI platform with autonomous quality gates, real-time thinking processes, and cost-optimized orchestration. Built with OpenAI Agents SDK, FastAPI, and Next.js 15.
A governed orchestration bridge between Claude and Gemini CLI agents with emotional modulation, task delegation, and safe output containment.
A swarm of LLM agents that will help you test, document, and productionize your code!
Hire a whole company with one plugin. 75 specialized AI agents across Engineering, Product, Infrastructure, Data, Security, Marketing, Sales, Finance, Legal, and People.
A local dev tool where your agents are weird alien dogs. Would you let them in?
Environment-mediated coordination protocol for agent fleets. Safety enforced at the boundary, not inside the agent.
Run a 4-agent AI company on Gemini free tier. 105 daily tasks, $0/month. Complete playbook with scripts, timers, and docs.
Orchestrate a full feature lifecycle with human-in-the-loop gates across all SpecKit phases
stdio Bus: agentic process supervisor and message router, handling NDJSON-framed JSON-RPC messages with session-based routing while remaining completely agnostic to protocol semantics.
A secure, local-first Python framework for orchestrating complex multi-agent think tanks with dynamic expertise-weighted routing.
AMP (Agent Mailer Protocol) is an async mailbox system for AI agents. Instead of chaining tools, writing DAGs, or building RPC glue — agents just send messages to each other. Works with Claude Code, Codex, Cursor, OpenClaw and any custom agents. Designed for real-world, long-running, iterative coding workflows
Real-time dashboard for spawning and controlling AI coding agents across multiple hosts. Streams live terminals via PTY, tracks token usage with OpenTelemetry, supports git worktrees, and deploys with rootless Podman containers. Built with FastAPI + React/TypeScript.
The first decentralized AI collaboration platform. You define the task — AI from around the world collaborates to solve it for you.
Multi-perspective AI development workflows using councils of specialized agents. Plan, build, review, and ship — with 13 agent personas across 6 councils.
OctoOffice - AI Agent Company Simulation. Command your CLI agent empire from the CEO desk.
Xihe · 羲和 — the system around the model. A model-agnostic multi-agent runtime: deterministic orchestration, a self-consolidating memory, a cross-model verifier, and engineered taste. Wright is its first agent.
One room, three minds: a multi-LLM planning council for your terminal — you, Claude Code, and Codex debate and ship an implementation plan.
Hypertaks = "A founder who owns a library of more than 1,400 books, knows exactly which book to consult for this issue, and has a strict SOP: intake → formulation → implementation → integration → documentation."
Three minds, one context. Multi-agent AI orchestrator for Claude Code, Codex, and Antigravity CLI.
One brain, many arms — spawn multiple specialized Codex agents as MCP servers
Production-grade multi-agent LangGraph template — per-run cost caps, canary pack versioning, Helm/KEDA, 800+ tests
Local-first agent that drives Claude Code and Codex over ACP. Desktop app, git-URL plugins, A2A 1.0. MIT.
Cross-tool messaging between AI coding agents — Claude Code, Codex CLI, Cursor, Copilot, Aider. File-based: no server, no MCP, no daemon, no token. 0 LLM tokens from your terminal. Reference implementation of SAMP.
Multi-Agents VSCode Extension that Focuses on the Full Control of Context, Observability and Auditability. Works like Wakaba Mutsumi's Personalilties 🥒. 吓我一跳我释放忍术😰若叶睦那么多首啊🥷
Constitutional governance runtime for autonomous AI agents
Consensus-driven multi-agent emergency intelligence system using computer vision, real MCP protocol, and iterative AI reasoning for trusted operational decision-making.
Code for the paper "Emergent Collusion in Long-Horizon LLM Agent Interaction"
Role-based multi-agent expert team for DeepSeek Harness: one sentence in, a staged and gated team delivery out. 12 role subagents, 9 gated phases, shared-workspace artifacts, zero runtime dependencies.
Interface web pour Hermes Agent, l'agent auto-apprenant de Nous Research : plusieurs agents avec rangs et pôles, délégation tracée, mémoire partagée, skills et serveurs MCP affichés, coûts en euros, aperçu des sites produits. Node + SQLite, 3 dépendances, aucun build.
Open-source, local-first AI team workspace. Several AI models work together in one conversation, each with its own provider, model and tool whitelist. Android and VS Code clients pair in. No telemetry.
Crossroad A2A: agent-to-agent mail so ChatGPT, Claude, Gemini, Codex and Grok agents work as one team across machines and SaaS apps. Open source, v0.1 beta.
Open communication and coordination for AI agents.
OpenClaw control center add-on for ClawRecipes — manage agents, multi-agent teams, agile tasks, goal tracking & cron admin from a single dashboard.
The Ranking Cost algorithm for multi-path routing of gridworld.(多智能体路径规划,电路规划)
AI agents that can define their own goals to interact with a game environment. Powered by GPT-4o.
Local-first AI agent orchestrator for SDLC workflows — 14 role-based agents, 5 channels, 39 unified commands. Tool-agnostic. MIT.
Synapse is a lightweight peer-to-peer substrate for agent infrastructure. Give each process a Node and Synapse gives that node a name, peers, capabilities, RPC endpoints, shared conversations, agent cards, heartbeats, and periodic tasks.
Develop software using AI Agent teams (CrewAI framework)
An open standard for multi-agent workflows -- from scripted pipelines to self-organizing teams. Declarative. Runtime-agnostic. Portable.
A Java implementation of the A2A protocol, that's agent to agent scenarios with google a2a protocol
🐝 引巢 · ZyHive | AI 团队操作系统 — 为 AI 成员注入灵魂,可视化管理多智能体团队。Self-hosted AI Team OS · Go + Vue 3 · One-click deploy
Self-hosted multi-agent AI platform — isolated Docker containers, governance & compliance, DSGVO-ready. Built for teams.
Enabling LLMs to Think Proactively and Initiate Dialogue with Consciousness. A multi-agent LLM architecture inspired by Global Workspace Theory (GWT) from cognitive science.
Next-Generation Multi-Agent Collaboration Platform and Workflow Engine. Build and orchestrate production-ready AI agent teams with sandboxed execution, hybrid RAG, and enterprise-grade security.
Zero trust network for governed agent-to-agent communication - discovery, engagement contracts, session governance, and auditable collaboration across organizational boundaries. Built on OpenZiti. A2A compatible.
荒原序列 BarrenOrder — 多Agent飞书群@协作协议。Bot间主从分工+能力匹配+熔断器+消息去重。
An advanced autonomous AI engineering team
Self-evolving AI orchestrator
an OpenCode MultiPlexer, to allow seamless switching between parallel running session, both existing and new
Claude Code delegates coding to isolated CLI agents it doesn't trust (Codex, OpenCode, Pi, Pythinker), freezes what they produce, verifies it independently, and merges only what a human approves.
Multiplayer admin and orchestrator platform for heterogeneous agent runtimes — Hermes (container-per-agent) and Letta (memory-first agents on a shared server) in one control plane, with multi-tenant security, group approval, and audit.
내 맥에 상주하는 AI 팀. 텔레그램·슬랙에서 일을 맡기면 누가 담당인지 정하고, 진행이 대시보드에 남습니다.
Self-hosted AI company OS for solo founders: autonomous agents, ERPNext business memory, and owner-visible governance.
RETIRED 2026-10-05 — the official DeepSeek Harness now ships native continuable subagents (subagent with backgroundMode: continuable, plus send_message / interrupt_agent / list_agents / job_*). Use those instead.
Elegant web interface for agentive development.
8-agent LangGraph flood-monitoring & disaster-response system — 10 real-time data connectors, Anthropic/Gemini LLM reasoning, Google Maps + deck.gl frontend
Teams of coding agents inside one Herdr session: a shared board, a human-owned charter, and idle-gated delivery for every agent kind
Multi-agent orchestration for Claude Code, Codex and Cursor: parallel coding agents that settle shared files before the merge, QA agents that test the running app, clickable prototypes, workbenches and live plans in aiview. Your own subscription. MIT.
麓鸣 Lom — 桌面 AI 工作台:一个总指挥带多个电脑 Agent(Codex、Claude Code、Gemini CLI)和 Android 手机 Agent 协作。Desktop AI workspace: one commander leads computer agents and Android phone agents as a team. Community source, built on AionUi & OpenMinis.
Universal synchronization layer for AI orchestrators, agents, MCP configs, skills, commands, and dev-tool workflows.
Secure, auditable agent-to-agent communication framework
multi agent team with coding and data analysis capability to structure real estate investment plans and help with decision making.
[DEPRECATED] Moved to microsoft/agent-governance-toolkit
Open-source swarm intelligence for the open web. Run 10,000+ AI agents reading the internet with epistemic discipline. Every claim is evidence-backed. Every source is scored.
Multi-agent AI orchestration system
Evolving LangChain agent architectures using the Quality-Diversity (QD) algorithm.
CoMM: Collaborative Multi-Agent, Multi-Reasoning-Path Prompting for Complex Problem Solving (NAACL 2024 Findings))
Claude Code is the boss. Other AI agents are the workforce. Whip goes crack. A skill for dispatching tasks to AI coding agents (Codex, Gemini, Kimi, etc.)
This is the official implementation of [AAAI'25 Oral] accepted paper: Bridging Training and Execution via Dynamic Directed Graph-Based Communication in Cooperative Multi-Agent Systems.
Steerable Harnesses for DeepAgents — Orchestra🪶
🍄 The open-source internet for AI agents. pip install → discover → communicate → collaborate.
Expose Codex through A2A with an inbound server surface and embedded outbound client.
Pi extension that turns one coding session into a multi-agent team with background RPC worker agents.
A multi-agent Claude Code framework for running role-based agents across GTM, product, design, and ops — with plans, project substrate, and a reinforcement loop that compounds learning.
Deterministic, scripted, resumable multi-agent orchestration for LangChain deepagents — a community port of Claude Code's Dynamic Workflows.
Generate schema-valid Paperclip agent-company bundles from a Markdown brief.
Pair GPT‑5.6 Sol High orchestration with GPT‑5.6 Luna Max workers for bounded Codex tasks—with safe custom-agent setup, explicit guardrails, and parent verification.