Open-Source AI Agent Directory

Browse and compare 30+ open-source AI agent projects. Find the right tool for autonomous coding, research, workflow automation, and more.

Why Open-Source AI Agents?

Open-source AI agents give you full control over your data, deployment, and customization. Unlike proprietary solutions, you can inspect the code, modify behaviors, and self-host on your own infrastructure.

This directory lists the most popular and actively maintained open-source AI agent projects in 2026. Each project is categorized by its primary use case: autonomous task execution, coding assistance, multi-agent orchestration, workflow automation, conversational AI, and more.

Whether you're building a personal AI assistant, automating DevOps workflows, or researching multi-agent systems, this directory helps you find and compare the right tools.

Filter by category

30 agents

OpenClaw

15k+

Local-first AI agent gateway with multi-platform support, skills marketplace, and privacy-first architecture.

TypeScriptApache 2.0Autonomous

AutoGPT

160k+

One of the first autonomous AI agents. Generates sub-tasks and executes them independently using GPT-4.

PythonMITAutonomous

BabyAGI

20k+

Minimalist task-driven autonomous agent. Pioneered the task creation and prioritization loop for AI agents.

PythonMITAutonomous

CrewAI

22k+

Framework for orchestrating role-playing autonomous AI agents. Build teams of agents that collaborate on tasks.

PythonMITMulti-Agent

LangGraph

8k+

Build stateful, multi-actor applications with LLMs. Graph-based agent workflow orchestration by LangChain.

PythonMITMulti-Agent

MetaGPT

45k+

Multi-agent framework that assigns different roles (PM, architect, engineer) to GPT models for software development.

PythonMITMulti-Agent

Microsoft AutoGen

35k+

Framework for building multi-agent conversational systems. Agents can chat with each other to solve complex tasks.

PythonMITMulti-Agent

ChatDev

26k+

Virtual software company with AI agents playing CEO, CTO, programmer, tester roles to build software collaboratively.

PythonApache 2.0Multi-Agent

CAMEL

6k+

Communicative Agents for "Mind" Exploration. Research framework for multi-agent cooperation and communication.

PythonApache 2.0Multi-Agent

OpenManus

12k+

Open-source AI agent framework focused on planning and tool use. Modular architecture for custom agent workflows.

PythonMITAutonomous

GPT Engineer

52k+

Generate entire codebases from a single prompt. AI agent specialized in code generation and project scaffolding.

PythonMITCoding

Sweep

7k+

AI-powered code review and bug fix agent. Automatically creates PRs to fix GitHub issues.

PythonAGPL-3.0Coding

Aider

25k+

AI pair programming in your terminal. Edit code in your local git repo with natural language.

PythonApache 2.0Coding

Continue

20k+

Open-source AI code assistant for VS Code and JetBrains. Autocomplete, chat, and edit with any LLM.

TypeScriptApache 2.0Coding

Open Interpreter

55k+

Let LLMs run code on your computer. Natural language interface to your computer's capabilities.

PythonAGPL-3.0Terminal

ShellGPT

10k+

Command-line productivity tool powered by AI. Generate shell commands, code snippets, and documentation.

PythonMITTerminal

Khoj

17k+

Self-hosted AI personal assistant. Search your notes, documents, and images using natural language.

PythonAGPL-3.0Assistant

Quivr

37k+

Your second brain powered by AI. Chat with your documents, notes, and data using any LLM.

PythonApache 2.0Assistant

PrivateGPT

54k+

Interact with your documents using LLMs. 100% private, no data leaves your execution environment.

PythonApache 2.0Assistant

Dify

55k+

Open-source LLM app development platform. Visual workflow builder with RAG, agent, and model management.

TypeScriptApache 2.0Platform

n8n

50k+

Workflow automation tool with 400+ integrations. Add AI agents to automate complex business processes.

TypeScriptSustainable UseWorkflow

Flowise

32k+

Drag & drop UI to build customized LLM flows. Low-code platform for AI agent pipelines.

TypeScriptApache 2.0Platform

Langroid

3k+

Harness LLMs with multi-agent programming. Lightweight framework for building LLM-powered applications.

PythonMITMulti-Agent

SuperAGI

15k+

Infrastructure for building, managing, and running autonomous AI agents at scale.

PythonMITAutonomous

AgentGPT

32k+

Deploy autonomous AI agents in your browser. Configure and run agents with a simple web interface.

TypeScriptGPL-3.0Autonomous

TaskWeaver

5k+

Microsoft's code-first agent framework. Converts user requests into executable code with plugin support.

PythonMITCoding

Botpress

13k+

Open-source platform for building conversational AI bots. Visual flow editor with NLU and integrations.

TypeScriptMITPlatform

Rasa

19k+

Open-source machine learning framework for text and voice-based AI assistants and chatbots.

PythonApache 2.0Platform

Haystack

18k+

End-to-end NLP framework for building AI applications with LLMs. Pipelines, RAG, and agent capabilities.

PythonApache 2.0Platform

Composio

12k+

Tooling infrastructure for AI agents. 150+ pre-built integrations for GitHub, Slack, Gmail, and more.

PythonAGPL-3.0Workflow

How to Choose the Right AI Agent

1

Define Your Use Case

Are you automating code reviews, building a chatbot, or orchestrating complex workflows? Start with your goal.

2

Check Privacy Requirements

If data privacy matters, choose self-hosted solutions like OpenClaw or PrivateGPT over cloud-only platforms.

3

Evaluate Community & Maintenance

Active GitHub repos with recent commits and responsive maintainers are more reliable long-term.

4

Consider Integration Needs

Check if the agent supports your existing tools: Slack, GitHub, databases, APIs, and deployment platforms.

Frequently Asked Questions

What is an open-source AI agent?

An open-source AI agent is a software program with publicly available source code that can autonomously perform tasks using AI models. You can inspect, modify, and self-host these agents on your own infrastructure.

Which AI agent is best for beginners?

OpenClaw and AgentGPT are great starting points. OpenClaw offers a simple installation process with a skills marketplace, while AgentGPT provides a browser-based interface requiring no setup.

Can I self-host all agents in this directory?

Most agents listed here can be self-hosted. Some (like Dify and n8n) also offer cloud-hosted versions. Check each project's GitHub README for deployment instructions.

How do multi-agent frameworks differ from single agents?

Single agents (AutoGPT, BabyAGI) work alone on tasks. Multi-agent frameworks (CrewAI, AutoGen, MetaGPT) coordinate multiple specialized agents that collaborate, similar to a team of experts.

Are open-source AI agents free to use?

The agent software itself is free. However, most agents require an LLM API key (OpenAI, Claude, etc.) which has per-token costs. Self-hosted open-source LLMs (Llama, Mistral) can reduce costs to just hardware expenses.

Related Resources

Start Building with OpenClaw

Install OpenClaw and run AI agents locally β€” private, free, and extensible.

Get Started