🤖 Awesome Agents

Awesome Agents is a curated list of open-source tools and products to build AI agents.

🆕 New list: Awesome ADE — open-source Agentic Development Environments: orchestrators for running coding agents in parallel, TUI and tmux harnesses, desktop and web workspaces, worktree and container isolation. Browse the list →

Table of Contents

• 🤖 Awesome Agents
  • Table of Contents
  • Frameworks
  • Testing and Evaluation
  • Software Development
  • Research
  • Conversational / General Agents
  • Game / Simulation
  • Memory - Knowledge Management
  • Automation
    • Browser

Frameworks

• OpenClaw: Open-source AI agent framework that turns LLMs into persistent, proactive personal AI agents with multi-channel messaging (Signal, Telegram, Discord, WhatsApp), cron scheduling, memory systems, MCP integration, skill plugins, sub-agent spawning, and browser automation.
• Hermes Agent: The agent that grows with you
• Transformers Agents: Provides a natural language API on top of transformers
• LlamaIndex: provides a central interface to connect your LLM's with external data.
• LangChain: the original 🐍
• Botpress: The building blocks for building chatbots
• Haystack: NLP framework to interact with your data using Transformer models and LLMs
• Semantic Kernel: Microsoft C# SDK to integrate cutting-edge LLM technology quickly and easily into your apps
• AGiXT: Dynamic AI agent automation platform that orchestrates instruction following and complex task management.
• LLM Agents: Build agents which are controlled by LLMs
• e2b: Open-source platform for building & deploying virtual developers’ agents
• Dust: Design and Deploy Large Language Model Apps
• MetaGPT: The Multi-Agent Meta Programming Framework: Given one line Requirement, return PRD, Design, Tasks, Repo and CI
• Lagent: A lightweight framework for building LLM-based agents
• Autogen: Enable Next-Gen Large Language Model Applications.
• AG2: AG2 (from AutoGen creators) is an open-source programming framework for building AI agents and facilitating cooperation among multiple agents to solve tasks.
• AgentVerse Provides a flexible framework that simplifies the process of building custom multi-agent environments for LLMs
• Maestro: A framework for Claude Opus to intelligently orchestrate subagents.
• AgentScope: Start building LLM-empowered multi-agent applications in an easier way.
• CrewAI: Cutting-edge framework for orchestrating role-playing, autonomous AI agents.
• Swarm: Educational framework exploring ergonomic, lightweight multi-agent orchestration by OpenAI.
• agency-swarm: Reliable agent framework built on top of the latest OpenAI Assistants API.
• Upsonic: Reliable agent framework that supports MCP.
• Mastra: Mastra is an opinionated TypeScript framework that helps you build AI applications and features quickly.
• Vectara-agentic: Vectara-agentic is a framework for creating AI Assistants and agents using Vectara.
• AgentDock: Stop wrestling with countless APIs and complex integrations. AgentDock provides the open-source foundation to build, manage, and deploy production-ready AI agents and workflows, frictionlessly.
• Swarms Framework Bleeding-Edge Multi-Agent Orchestration Framework for Enterprise Applications
• Strands Agents SDK: A model-driven approach to building AI agents in just a few lines of code.
• VoltAgent - An open source TypeScript Framework for building AI agents with built-in LLM observability.
• Agentic Context Engine: Self-improving agents that learn from execution feedback. LangChain integration for agents that curate their own context.
• Astron: Enterprise-grade, commercial-friendly agentic workflow platform for building next-generation SuperAgents.
• Ailoy: A comprehensive framework for building AI agents that run anywhere, with full local-AI and WASM support.
• PraisonAI: Production-ready Multi-AI Agents framework with self-reflection. Fastest agent instantiation (3.77μs), 100+ LLM support, MCP integration, workflows, memory, and both Python & JavaScript SDKs.
• AgentField: Open-source infrastructure for AI backends. Kubernetes for orchestration, Okta for identity—production-ready from day one.
• Cortex Memory: A complete solution for agent memory, from extraction and vector search to automated optimization, with a REST API, MCP, CLI, and insights dashboard out-of-the-box.
• Pipecat: Open Source framework for voice and multimodal conversational AI.
• LoongFlow: A Evolve Agent Development Framework. From atomic components and development frameworks to core scenario Agents
• Agentset: Open-source production-ready RAG platform with built-in agentic reasoning, hybrid search, and multimodal support.
• Pilot Protocol: An overlay network stack that gives AI agents virtual addresses, encrypted UDP tunnels, NAT traversal, and mutual trust. Zero dependencies, written in Go.
• hcom: Let AI agents message, watch, and spawn each other across terminals. Claude Code, Gemini CLI, Codex, OpenCode.
• nanobot: Ultra-lightweight personal AI assistant framework (~4,000 lines of Python). Supports MCP, 9+ chat channels, and extensible skills system.
• Hive: Open-source AI agent framework for building goal-driven, self-improving autonomous agents. Define outcomes in natural language and auto-generate agent graphs with built-in evolution loops, MCP integration, and 100+ tools.
• ConnectOnion: Simple Python framework for production-ready AI agents — 2-line creation, functions as tools, 12 lifecycle hooks, plugin system, multi-agent networking with trust
• SwarmClaw: Self-hosted AI runtime for multi-agent orchestration with heartbeats, schedules, delegation, memory, and runtime skills across OpenClaw, Claude Code, Codex, Gemini CLI, and major LLM providers. Ships as a desktop app (Electron) and CLI.
• smolagents: Open-source Python library designed to make it extremely easy to build and run agents using just a few lines of code.
• Open Multi-Agent: TypeScript-native multi-agent orchestration that plans task DAGs at runtime in your own environment, with approval gates, tracing, evaluation, checkpoints, and resume support. Requires Node.js 20+.
• Aeon: Autonomous agent framework that runs unattended via GitHub Actions, with self-healing skills, output quality scoring, persistent memory, and reactive triggers.
• Cordum: Out-of-process governance control plane for autonomous AI agents — pre-dispatch policy enforcement, approval gates, and signed audit trails for production governance.
• trpc-agent-go: Go framework for building production agent systems with graph workflows, tools, memory, A2A, AG-UI, MCP, evaluation, and observability.
• Agentlas-OS: Local-first, model-agnostic Agent Operation Environment (AOE) for building, borrowing, and orchestrating specialist agents and teams across supported LLM hosts, with owner-scoped packages and host-enforced permissions.
• Lobu: Open-source context and control plane for AI agents. Connect company tools once and share permission-aware context across ChatGPT, Claude, Codex, OpenClaw and custom agents, with MCP, scoped identities, approvals and audit.
• OpenProgram: Self-programming AI agent framework where agents create and refine their own workflows, with multi-agent orchestration, persistent sessions, sandboxed execution, and desktop and web interfaces.
• SandBase Harness: Local-first managed agent runtime with persistent sessions, sandboxed execution, MCP tools, audit, replay, and a human console.
• Agent Swarm: Self-hosted multi-agent system where a lead agent delegates tasks to specialized workers with shared memory, tools, schedules, and review gates.
• agent-sdk-go: Framework for building stateful AI agents in Go.
• OpenHuman: Local-first personal AI agent for the desktop: agents, workflows, MCP client, local models via Ollama.
• Tenuo: Per-call authorization for agent tools. Each call is checked at the argument level, and authority is traced across delegation chains.
• Pydantic Deep Agents: Python agent framework built on Pydantic AI, with planning, sub-agents, file and shell tools, MCP and optional Docker sandboxing; also ships as a terminal coding agent.

Testing and Evaluation

• Voice Lab: A comprehensive testing and evaluation framework for voice agents across language models, prompts, and agent personas.
• Open-RAG-Eval: an open source RAG evaluation framework that does not require golden answers, and can be used to evaluate performance of RAG tools connected to an AI Agent (Agentic RAG)
• EvoAgentX: EvoAgentX is building a Self-Evolving Ecosystem of AI Agents, it will give you automated framework for evaluating and evolving agentic workflows.
• Arize-Phoenix: Arize-Phoenix is an open source library for agent testing, evaluation and observability.
• Manifest: Open-source, real-time cost observability platform for AI agents. Track tokens, costs, messages, and model usage with a local-first dashboard. Supports 28+ LLM models, OTLP ingestion, self-hosted.
• agent-qa: Self-improving agentic QA harness for web and mobile tests. Write tests in natural language, use memory to adapt to UI changes, and catch regressions before releases ship.
• Latitude: Open-source LLM observability and evaluation platform for AI agents. Traces agent runs, clusters production failures into issues, and auto-generates evals from real-world failures. Built on OpenTelemetry, self-hostable.
• ClawMetry: Open-source, zero-config real-time observability dashboard for AI agent runtimes (Claude Code, Codex, Cursor, Hermes and more) with sessions, token and cost analytics, and OTLP ingestion.

Software Development

• MetaGPT: The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming
• OpenHands: 🙌 OpenHands: Code Less, Make More. (formerly OpenDevin), a platform for software development agents powered by AI
• GPT Pilot: GPT Pilot is the core technology for the Pythagora VS Code extension that aims to provide the first real AI developer companion.
• Aider: aider is AI pair programming in your terminal
• Devika: Devika is an Agentic AI Software Engineer that can understand high-level human instructions, break them down into steps, research relevant information, and write code to achieve the given objective
• RepoAgent: An LLM-powered repository agent designed to assist developers and teams in generating documentation and understanding repositories quickly
• DSPy: The framework for programming—not prompting—foundation models
• ThinkGPT: Agent techniques to augment your LLM and push it beyond its limits
• PyCodeAGI: A small AGI experiment to generate a Python app given what app the user wants to build
• SuperAGI: SuperAGI - A dev-first open source autonomous AI agent framework
• Plandex: An AI coding engine for complex tasks
• Codel: ✨ Fully autonomous AI Agent that can perform complicated tasks and projects using terminal, browser, and editor.
• DB GPT: Interact your data and environment using the local GPT, no data leaks, 100% privately, 100% security
• Agency: 🕵️‍♂️ Library designed for developers eager to explore the potential of Large Language Models (LLMs) and other generative AI through a clean, effective, and Go-idiomatic approach
• MicroAgent: Agents Capable of Self-Editing Their Prompts / Python Code
• SWE Agent: SWE-agent takes a GitHub issue and tries to automatically fix it, using GPT-4, or your LM of choice.
• AgentRun: The easiest, and fastest way to run AI-generated Python code safely
• Claude Engineer: Claude Engineer is an interactive command-line interface (CLI) that leverages the power of Anthropic's Claude-3.5-Sonnet model to assist with software development tasks.
• Vision agent: Vision Agent is a library that helps you utilize agent frameworks to generate code to solve your vision task.
• TypedAI: TypeScript AI agent platform with Autonomous agents, Software developer agents, AI code review agents and more.
• Cline: Open-source AI coding agent giving developers direct access to frontier models with full transparency.
• OpenCode: The AI coding agent built for the terminal.
• Stakpak: Open-source DevOps agent to help you secure, deploy, and maintain production-ready infrastructure.
• ReviewCerberus - 100% free, open-source AI code review tool for analyzing git branch differences with comprehensive security, performance, and quality analysis.
• Dorothy: Open-source desktop app to orchestrate multiple AI CLI agents (Claude Code, Codex, Gemini) simultaneously with automations, Kanban management, remote control, and MCP servers.
• Greywall: Deny-by-default command sandbox for AI coding agents with filesystem isolation, network control via transparent proxy, built-in profiles for agents like Claude Code or OpenCode, and learning mode for auto-generating configs.
• Maestro Orchestrate: Multi-agent development orchestration platform coordinating 22 specialized AI agents through 4-phase workflows with native parallel execution, persistent sessions, and least-privilege security tiers. Runs on Gemini CLI, Claude Code, and Codex.
• amux: Agent multiplexer for running dozens of parallel Claude Code sessions with a web dashboard, self-healing watchdog, kanban board, agent-to-agent REST API, and mobile PWA. Python 3 + tmux.
• AgentsMesh: The AI Agent Workforce Platform. Remote AI workstations (AgentPods) with PTY sandbox + git worktree isolation, multi-agent collaboration via channels and pod bindings, built-in Kanban with MR/PR integration.
• KubeStellar Console: Multi-cluster Kubernetes dashboard with AI-powered operations agent (kc-agent) providing MCP tools for managing workloads across edge and cloud clusters. CNCF project.
• Bernstein: Open-source governance layer for AI agents.
• ctop: htop for AI coding agents. Monitor Claude Code and Codex CLI sessions with real-time CPU, memory, token usage, context window tracking, and cost estimates. Zero dependencies, pure Node.js.
• Frontman: Coding agent that lives in your browser, hooks into your dev server, sees the live DOM, component tree, CSS, routes, and logs, then edits the actual source files with instant hot reload.
• Maestro: Desktop command center for running multiple AI coding agents (Claude Code, Codex, Gemini CLI, etc.) in parallel, with Cue event automation, Auto Run playbooks, Group Chat across local and remote agents, and a maestro-cli that agents can drive themselves.
• Cate: Desktop IDE on an infinite zoomable canvas. Editors, terminals, browsers, and Claude Code agent panels float in spatial workspace instead of tabs; panels can dock or detach into separate windows. Electron + React + TypeScript, layout persists per project.
• Nanocoder: Local-first CLI coding agent built with React and Ink.
• zeroshot: CLI that orchestrates a planner, an implementer, and independent validators in isolated environments
• h5i: CLI that runs several coding agents on the same task, each in its own isolated git worktree sandbox, has them peer-review each other, then a neutral verifier replays every candidate, runs the tests, and merges the one that passes.
• Paperclip: Open-source platform to run coding agents as managed workers with tasks, approvals, budgets, and workspaces.
• Agent Teams AI: Open-source desktop orchestrator for autonomous coding-agent teams with task delegation, inter-agent messaging, Kanban, and code review across multiple agent runtimes.
• Tura: A local, open-source coding agent for developers who are tired of vague skill claims, token-saving extensions with no evidence, and agents that change a repository before understanding it.
• Orca: A DeepSeek-native terminal coding agent in Rust with OS-level sandboxing, persistent goal mode, background tasks, and verification gates
• fractal: Hierarchical agent loops with recursive self-organization.
• OpenCodeReview: An open-source hybrid code reviewer combining deterministic rules with an LLM agent (tool-use, line-level comments); self-hostable, bring-your-own model.
• OpenBitFun: Open-source desktop agent with a Rust runtime for real repositories, browser, terminal, and desktop execution, extensible through MCP, Skills, and custom agents.
• CompozyOS: Open-source operating system for AI agents — plug in the agent CLIs you already use and run them as a team on loops and schedules, with shared memory, permissions and approvals
• Ouroboros (Agent OS): A Socratic interview gates the spec on an ambiguity score, then one command drives execution, a staged evaluation gate, and a budgeted evolution loop that runs the semantic stage only.
• Atomic Agent: Local-first CLI and TUI coding agent that runs open-weight models entirely on your machine via a llama.cpp fork.
• Keen Code: Open-source, context-aware terminal coding agent written in Go with multiple providers, Turn Memory for controllable cross-turn tool-output retention, skill-driven MCP integration, subagents, Agent Skills, and hashline edits.
• Ordewell: Terminal CLI and TUI that turns one goal into an ordered, editable plan of coding agent tasks, then runs each task as its own session on the harness you pick per task, with its own model, mode and effort, and accepts a task as done only when its unique completion marker appears in that runner's output.
• Orbi: Self-hosted autonomous coding agent that takes a labelled GitHub Issue through implementation, an independent review session, merge of the exact reviewed head, and a tagged release.
• agent-manager: Terminal UI that runs coding-agent CLIs side by side, each in its own persistent tmux session, with live status, quick prompts, git worktrees and diff review.
• mu: Coding agent built on Pi, with a CLI and a desktop app, where a small, fast judge model (Jev, a local judge or any LLM) makes the routine calls at more than 30 decision points, such as which chunks of a long tool output enter the context, whether a rule-flagged command was asked for, and whether a fetched page or MCP result carries instructions aimed at the model; every verdict is logged locally.
• wmux: Desktop workspace for running Claude Code, Codex, Gemini CLI and other CLI agents side by side on Windows and macOS, with up to 8 parallel tasks each in its own git worktree, hunk-level diff adoption, approval gates for agent tool calls, agent-to-agent channels and sessions that survive reboots.
• ReevesAgents: Local workspace that runs coding-agent CLIs side by side in tmux, driven from a TUI, a local Web UI, or the CLI, with an opt-in MCP server so one agent can spawn, read, and steer the others.

Research

• GPT Researcher: GPT Researcher is an autonomous agent designed for comprehensive online research on a variety of tasks.
• BlockAGI: BlockAGI conducts iterative, domain-specific research, and outputs detailed narrative reports to showcase its findings
• data-to-paper: data-to-paper: AI-driven research from data to human-verifiable research papers
• AI Scientist: The AI Scientist: Towards Fully Automated Open-Ended Scientific
• Storm: An LLM-powered knowledge curation system that researches a topic and generates a full-length report with citations.
• OpenLens AI: Fully Autonomous Research Agent for Health Infomatics
• DeepAnalyze: first agentic LLM for autonomous data science, supporting specific data tasks and data-oriented deep research (produce analyst-grade research reports).
• AIDE: AI-Driven Exploration — ML engineering agent that uses tree search to automate experiment design, code generation, and evaluation against any metric.
• Cynative: Deep cybersecurity research agent for your cloud, code and runtime. Read-only, sandboxed.
• Kapso: Long-running agents that optimize AI and Data systems, and learn from every experience.
• Jev Social: Open-source social-media research agent where Jev selects bounded read-only operations across Instagram, TikTok, and LinkedIn; the local socai CLI collects browser evidence and the app streams posts, media, and a cited report.
• Research Toolkit: Research methods, workflow tools, and independent review for AI agents producing source-backed reports.

Conversational / General Agents

• Colossal-AI: LLM training and inference system whose ColossalChat application implements LLM with RLHF
• RasaGPT: RasaGPT is the first headless LLM chatbot platform built on top of Rasa and Langchain.
• SuperAgent: Deploy LLM Agents to production
• ix: Autonomous GPT-4 agent platform
• Multi-Modal LangChain agents in Production: Deploy LangChain Agents and connect them to Telegram
• Autonomous HR Chatbot: An autonomous agent that can answer HR related queries autonomously using the tools it has on hand
• LLama Cpp Agent: The llama-cpp-agent framework is a tool designed for easy interaction with Large Language Models
• Letta: Platform for stateful agents with advanced memory that learn and self-improve 📚🦙
• joinly: Voice-first AI Assistant for online meetings that can actively participate and solve tasks live during the meeting
• Gobii: Gobii is an open-source platform for deploying and managing browser-use agents at scale with a conversational interface and API
• ClaudeClaw: Persistent agent orchestrator plugin for Claude Code — multi-channel routing (Slack, WhatsApp, Telegram), OS-level sandbox isolation, composable extensions, structured memory, webhook triggers
• OpenAgent: Next-generation personal AI assistant powered by LLM, RAG and agent loops. Supports computer-use, browser-use, coding agent, 30+ model providers, visual workflow builder, and MCP tools. Self-hostable with admin dashboard.
• MateClaw: Open-source personal AI operating system on Spring Boot + Spring AI Alibaba — one JAR ships a web admin, desktop app (bundled JRE 21), embeddable widget, and 8 IM channels sharing the same agent.
• Orkas: Local-first multi-agent desktop application where a Commander coordinates specialist agents for research, coding, data analysis, documents, and media; MIT-licensed with bring-your-own model keys.
• Ouroboros: Self-hosted general-purpose agent with durable identity and memory, reviewed self-modification, specialist subagent swarms, and desktop or headless operation.
• 5dive: Run a company of AI agents on a server you own. Named agents on an org chart with a shared backlog, handing off work, and you decide the rest on your phone.
• Raven: The Harness of Harnesses, built for recursive self-improvement: a trusted, persistent, self-evolving multi-agent ecosystem for all-domain collaboration.

Game / Simulation

• SkyAGI: Emerging human-behavior simulation capability in LLM agents
• Voyager: An Open-Ended Embodied Agent with Large Language Models
• Enclave: Self-hosted single-owner AI social world; each instance is populated by autonomous AI residents with personalities, schedules and relationships who chat, form group conversations, post to a social feed and proactively message the owner.
• MiroShark: Universal swarm-intelligence engine — drop in a scenario and hundreds of grounded LLM agents simulate Twitter, Reddit, and a prediction market hour-by-hour, with counterfactual branching, per-agent MCP tools, and a public gallery of finished runs.

Memory - Knowledge Management

• Private GPT: Interact privately with your documents using the power of GPT, 100% privately, no data leaks

• Local GPT: Inspired on Private GPT with the GPT4ALL model replaced with the Vicuna-7B model and using the InstructorEmbeddings instead of LlamaEmbeddings

• Second Brain AI Agent: A streamlit app to dialog with your second brain notes using OpenAI and ChromaDB locally.

• SAGE: Institutional memory for AI agents — every memory goes through BFT consensus before it's committed. 4 application validators, 13 MCP tools, runs locally.

• Hindsight: State-of-the-art long-term memory for AI agents by Vectorize.

• piia-engram: Cross-tool persistent memory for AI agents. Local-first identity and knowledge layer and any MCP-compatible client.

• Caura: Open-source governed shared memory for AI agent fleets with cross-agent recall, permissions, audit trails, and persistent memory.

• Statewave: Open-source memory runtime for AI agents that transforms events into structured memories, enabling memory evolution, consolidation, supersession, and long-term context management across agent workflows.

• CodeAlmanac: Self-updating repo wiki for AI coding agents that tracks project conversations, decisions, and context locally inside the repository.

• IWE: Markdown knowledge graph for you and your AI agents — editor LSP plus CLI and MCP server so agents can search, retrieve, and refactor plain-text notes.

• OpenViking: Self-evolving Context Database for AI Agents — unifies Agent Memory, Knowledge RAG, and Skills behind a single interface, with a portable Agent Plugins package format.

• pond: Lossless storage and search for AI agent sessions across every agentic client.

• Busabase: Local-first database and workspace for AI agents that keeps structured data, durable knowledge, documents, files, skills, and apps in one searchable place. Agents connect through MCP, OpenAPI, CLI, or coding-agent skills, and material writes can remain reviewable ChangeRequests before becoming canonical.

Automation

• DemoGPT: DemoGPT enables you to create quick demos by just using prompt. It applies ToT approach on Langchain documentation tree.
• RestGPT: An LLM-based autonomous agent controlling real-world applications via RESTful APIs
• XAgent: An Autonomous LLM Agent for Complex Task Solving
• LLM Agents: Build agents which are controlled by LLMs
• uAgents: A fast and lightweight framework for creating decentralized agents with ease.
• Maestro: A framework for Claude Opus to intelligently orchestrate subagents.
• llama-agents: llama-agents is an async-first framework for building, iterating, and productionizing multi-agent systems, including multi-agent communication, distributed tool execution, human-in-the-loop, and more
• Agno: Build, run, and manage agent platforms.
• AgentK: An autoagentic AGI that is self-evolving and modular.
• ADAS: Automated Design of Agentic Systems
• Giselle: Giselle is an agentic workflow builder that empowers you to create AI-driven solutions with ease.

Browser

• Skyvern: Automates browser workflows with LLMs and computer vision; its agents log in to portals, fill out forms, download files, and extract structured data.
• OpenAgents: An Open Platform for Language Agents in the Wild
• Steel Browser: Open-source browser infrastructure for AI agents with session-backed automation, extraction, screenshots/PDFs, an AI-native CLI, and a reusable steel-browser skill.
• Actionbook: Parallel Action CLI for AI agents. Run 50 actions across 20 sites at once.
• invisible-playwright: Playwright wrapper for a stealth-patched Firefox 150 binary. Drop-in replacement returning native Playwright Browser objects, fingerprint spoofing in the C++ source.
• Webcmd: Self-learning browser infrastructure for AI agents that records how a site is navigated, then compiles it into deterministic per-site CLI commands.
• socai: Open-source browser-use agent for social-media research that reuses a signed-in Chrome session to search and read posts and comments, capture media, and return structured evidence through CLI, TUI, or Desktop.