Multi Agent Systems GitHub Repositories (2026) · Open Source Tools

CAIPE is an open-source AI platform for building, governing, and operating AI agents and agentic workflows for platform engineering and beyond.

Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems

An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.

Give AI agents ambitious work without losing the plot. nac is an open-source harness for long-running tasks, using a central orchestrator, threads, and structured episodes to stay aligned with your intent.

Multi-Agent + Automation Harness: Graph-based Workflows, MCP, Self-Improving Agents. NextJs+React

Multi-agent orchestration for AI coding CLIs — Claude Code, Kiro, Codex, and more, coordinated in isolated tmux sessions

Ruby AI Agents SDK - Inspired by OpenAI Agents SDK

Agent OS: keep specialist agents in a hub, spin up a temporary orchestrator per task. Local-first, works with any model.

An agent runtime for Python. Give your agent real work. Keep control: every action checked before it runs, every run survives a crash without silently redoing anything, every step on the record.

An Aspire reference architecture for cloud-native microservices and agentic AI with MCP, A2A & AG-UI

Caura (formerly MemClaw) — governed shared memory for AI agent fleets. Multi-agent, multi-tenant, MCP-native. Trust tiers, keystone policies, audit trails, knowledge graph, self-improving retrieval. Apache 2.0.

the composable multi-agent shell

Foam-Agent: A large language model-based multi-agent framework for automating computational fluid dynamics workflows. Published in Computer Methods in Applied Mechanics and Engineering (CMAME).

Chat application with multi-agents system supports multi-models and MCP

JarvisCore is a runtime where AI agents operate as a fleet of equal peers. Agents discover one another by capability, execute tasks from a shared ledger, and authenticate to every external service through a zero-trust broker, never with their own keys.

Neuro AI System of Agent Networks

Universal AI Agent Swarm & MCP Plugin with 176 specialized skills and 8 autonomous subagents for Antigravity, Claude Code, Cursor, Windsurf, and OpenAI Codex.

Autopus-ADK is of the agents, by the agents. for the agents. Multi-model orchestration (consensus/pipeline/debate/fastest). Architecture-as-Code, Lore decision tracking, SPEC/EARS engine.

LLM agents: curated papers on planning, memory, tool use, multi-agent systems, evaluation, and safety.

One IDE for you and your AI team. A VS Code fork where Claude Code, Codex, Antigravity, and Muse Code work in parallel across multiple projects and features.