Agents

Flexible, lightweight open-source framework for orchestrating multiple AI agents to handle complex conversations with parallel execution capabilities. Apache 2.0 licensed

Agent2Agent (A2A) open protocol enabling communication and interoperability between opaque agentic applications. Donated to Linux Foundation by Google with 50+ technology partners. Apache 2.0 licensed

Deployable starter kit for building virtual towns where AI characters live, chat and socialize. Inspired by Stanford's Generative Agents research with persistent agent memory and social interactions. MIT licensed

Workshop-level automated scientific discovery via agentic tree search. Generates novel research ideas, runs experiments, and writes papers

Open-source automation platform with AI agents, MCP integrations, and self-hosted workflow orchestration

Library to build and auto-optimize LLM applications with LLM-AutoDiff for fine-tuning-free optimization. End-to-end workflow optimization with tracing and human-in-the-loop capabilities. MIT licensed

Reliable multi-agent orchestration framework built on top of the OpenAI Assistants API with organizational structure modeling

Code-first Python toolkit for building sophisticated AI agents with multi-agent orchestration, built-in evaluation, and flexible deployment. Model-agnostic with tight Google ecosystem integration. Apache 2.0 licensed

Open file format (.af) for serializing stateful AI agents with persistent memory and behavior. Share, checkpoint, and version control agents across compatible frameworks. Apache 2.0 licensed

Next-generation proxy and routing layer for AI agents and MCP servers, with Kubernetes-native transport, protocol interoperability, and service-mesh-style isolation for reliable agent infrastructure. Apache 2.0 licensed

Policy, safety, and execution controls for autonomous AI agents, including governance guardrails, sandboxing, and reliability checks. Apache 2.0 licensed

Flexible multi-agent orchestration framework with intelligent intent classification and context management. Supports Python and TypeScript with pre-built agents for Bedrock, Lex, and custom integrations. Apache 2.0 licensed

Open agentic framework that uses computers like a human. SOTA on OSWorld benchmark (72.6%) for GUI automation and computer control

Alibaba's production-ready multi-agent framework with 23K+ stars. Features built-in MCP and A2A support, message hub for flexible orchestration, and AgentScope Runtime for production deployment

Build, run, and manage agentic software at scale. High-performance framework for multi-agent systems with memory, knowledge, and tools

Command-line pair-programming agent

Workflow engine for deterministic AI coding agents. Define development processes as YAML workflows (planning → implementation → validation → review → PR) with isolated Git worktrees for parallel execution. MIT licensed

The original autonomous AI agent framework that sparked the agent revolution. Vision of accessible AI for everyone with modular agent architecture, benchmark testing, and forge-based agent building. 183k+ stars

Flexible multi-agent conversation framework

Pioneering task-driven autonomous agent that inspired the AI agent movement. Simple, elegant implementation of an AI agent that creates, prioritizes, and executes tasks autonomously. 22k+ stars

Production-ready multi-agent framework in Python and TypeScript. Features workflow orchestration, ACP/MCP protocol support, and deep watsonx integration. Part of Linux Foundation AI & Data program

Makes websites accessible for AI agents. Enables autonomous web automation, data extraction, and task completion with natural language instructions. MIT licensed

Apache incubating framework for building stateful AI applications (chatbots, agents, simulations). Monitor, trace, persist, and execute on your own infrastructure with built-in UI and pluggable memory. Apache 2.0 licensed

First and best multi-agent framework for building scalable agent systems. Apache 2.0 licensed with extensive tooling for agent communication and task automation

Multi-agent software development framework where AI agents collaborate as programmers, designers, and testers to build software. Apache 2.0 licensed

Pre-indexed local code knowledge graph for Claude Code, Codex, Gemini, and other coding agents to reduce context churn and token spend in developer workflows

Local pre-indexed code knowledge graph for coding agents to reduce token usage and redundant tool calls across Claude, Codex, and other agents

Tool integration layer for AI agents with 1000+ toolkits, authentication management, and sandboxed workbench. Powers tool use across major frameworks

Agentic orchestrator for parallel coding agents. Plans tasks, spawns agents, and autonomously handles CI fixes, merge conflicts, and code reviews. MIT licensed

Event-driven agentic orchestration platform providing durable and resilient execution engine for applications and AI agents. Battle-tested at Netflix, Tesla, LinkedIn, and J.P. Morgan with 30K+ stars. Apache 2.0 licensed

Role-based agent framework

Framework for programming language model pipelines with modules, optimizers, and evaluation loops

Hierarchical multi-agent system for deep research tasks with automated task decomposition and execution across complex domains

Open-source long-horizon SuperAgent harness that researches, codes, and creates. Handles tasks from minutes to hours with sandboxes, memories, tools, skills, subagents, and message gateway

Production-ready agentic workflow platform

Performance-oriented agent harness for coding agents with skills, memory, and security-aware orchestration across Claude Code, Codex, and more

MCP server for persistent AI agent memory with atomic notes, semantic linking, and SQLite or PostgreSQL storage

Open-source AI agent that brings Gemini's power directly into your terminal. Supports code generation, shell execution, and file editing with full Apache 2.0 licensing

Extensible on-machine AI agent for development tasks

Local-first code knowledge graph and code intelligence engine built in Golang with multi-repository support and real-time graph actualisation. Built for AI coding agents aiming to expose only the needed information, reducing token usage by up to 50x. Works natively with Claude, Codex, Hermes, and other agents

Modular Python framework for AI agents and workflows with chain-of-thought reasoning, tools, and memory. Enforces structures like sequential pipelines and DAG-based workflows for predictable AI systems. Apache 2.0 licensed

Efficient programming paradigm for steering language models. Control output structure with loops, conditionals, and regex constraints inline. Reduces latency and cost vs conventional prompting. MIT licensed

MCP-powered coding-focused cybersecurity agent framework for automated pentesting and bug-hunting workflows

Context compression proxy for tool outputs, logs, and RAG chunks, reducing token pressure while preserving intent for AI agents

Token and output compression middleware for AI agents and RAG workflows, reducing context footprint while preserving semantics (claims up to 60-95% fewer tokens). Apache 2.0 licensed

The agent that grows with you. Autonomous server-side agent with persistent memory that learns and improves over time

State-of-the-art long-term memory for AI agents by Vectorize. Fully self-hosted, MIT-licensed, with integrations for LangChain, CrewAI, LlamaIndex, Vercel AI SDK, and more

Production-grade multi-agent orchestration framework with 10K+ stars. Apache 2.0 licensed