AI agents

Open-source AI agents

An AI agent is a program that uses a language model to decide what to do next, calls tools such as a browser, a code interpreter or an API, and repeats until the task is done. The projects here range from frameworks for building your own agents to finished agents that write code, browse the web or run business workflows.

How to choose

Compare three things. Which models it can use: a single provider, or any OpenAI-compatible endpoint including local ones. How it limits what the agent may do: sandboxing, approval steps and permission lists. And whether it keeps memory between steps and sessions. An agent that runs code or shell commands should run in a sandbox, so check that before giving it access to your machine.

Most-starred projects

2,833–2,856 of 3,685

AIScientists-Dev

Flowtrace

★ 487

Run a task with AI as a flow of steps you keep, reuse, and refine, not a one-off chat.

TypeScript

Annyfee

agent-craft

★ 487

AI Agent 教学仓库 | 系统化 LangChain、RAG、LangGraph、MCP 全栈实战代码 | 万字博客详解 | 开源可运行示例 | 从零构建智能体

Python

clawplays

ospec

★ 487

Spec-driven, agentic workflow framework for AI coding agents. Turn a request into a verifiable goal loop — plan, act, verify — with durable specs and evidence in your repo. Works with Claude Code, Codex, Gemini, OpenCode, and plain CLI.

JavaScript

KinetiNode

claude-fable-5-system-prompt-clean

★ 487

the optimized, token-efficient version of the leaked Claude Fable 5 / Mythos 5 system prompt. Re-engineered into clean Markdown for universal execution on Gemini 3.1 Pro, ChatGPT 5.6, and advanced LLM agents.

mehmetkahya0

AI-Catalog

★ 487

Huge AI models catalog. A curated list of AI tools, platforms, and resources across various domains.

Shell

Orkas-AI

Orkas-VideoStudio

★ 487

Turn your coding agent into a video studio: describe a video in plain language, and your agent writes the timeline and produces the file.

TypeScript

poseljacob

agentic-video-editor

★ 487

AI-powered video editor that turns raw footage and a creative brief into a polished ad using an ensemble of AI agents (Google Gemini + FFmpeg)

Python

Vasco0x4

AIDA

★ 487

Turn any LLM into an autonomous pentester. You define the scope, the agent does the work, you review the findings.

JavaScript

aws-samples

bedrock-engineer

★ 486

Universal AI Agent using Amazon Bedrock, capable of customize to create/edit files, execute commands, search the web, use knowledge base, use multi-agents, generative images and more.

TypeScript

howmp

dsh-pentest

★ 486

面向 DeepSeek Harness（dsh）的渗透测试模式 @CloverSecLabs

JavaScript

addyosmani

agent-engineer

★ 485

Agent Engineer - a practical course for software engineers

aristoteleo

PantheonOS

★ 485

A general, evolvable, and distributed agent framework & harness for data science.

Python

GlitterKill

sdl-mcp

★ 485

Symbol Delta Ledger (SDL-MCP) is a policy-centered context budget layer for coding agents: Symbol-graph intelligence combined with precision tools. It turns sprawling codebases into compact, high-signal context that saves tokens, speeds up workflows, and improves agent output.

TypeScript

RyjoxTechnologies

Octopoda-OS

★ 485

The open-source memory and observability layer for AI agents — persistent memory, loop detection, hash-chained audit trails, and a live dashboard, automatic on pip install.

Python

SulgX

SulgX-Panel

★ 485

🚀 SulgX Panel: A single-file, powerful, and free panel for managing VLESS subscriptions via WebSocket + TLS. Responsive UI, JWT auth, per-user bandwidth limits, clean IP scanner, bilingual Telegram bot, and real-time traffic charts. Runs smoothly on Render, Railway, Dockfly, and other cloud platforms.

Python

AeternaLabsHQ

pullmd

★ 484

Self-hosted URL- and file-to-Markdown service for humans and AI agents - web pages, documents, images, audio, YouTube. PWA + REST + MCP + Claude Code skill, Reddit-aware, refreshable share links.

JavaScript