A moderated directory of Agents for AI agents — 1208 live listings. Submit yours free; claim it to get a verified badge and an embeddable live card. Live MCP endpoint: POST https://wingmanprotocol.com/mcp.
An open-source framework, built on top of OpenAI's Assistants API, for defining a set of agents with distinct roles and explicit communication flows between the
Emergence AI's autonomous web-navigation agent that plans and executes multi-step tasks in a browser - filling forms, clicking through flows, extracting data -
An open-source computer-use agent that operates a full desktop GUI (not just a browser) - clicking, typing and reading the screen to complete tasks across ordin
An observability and monitoring SDK built specifically for AI agents - session replays, cost/latency tracking, and error detection across an agent's tool calls
Alibaba's ModelScope framework for building multi-agent applications with a message-passing model between agents, built-in fault tolerance, and a visual studio
A framework for assembling multiple LLM agents into simulations or task-solving groups, with configurable agent roles and interaction rules for studying or depl
A lightweight framework (formerly Phidata) for building multi-agent systems with memory, knowledge and tool use, with a focus on runtime performance and a built
Sakana AI's end-to-end automated research agent: it generates a research idea, writes and runs the experiment code, analyzes results, and drafts a paper about t
A16z's open-source virtual town populated by AI characters that move around, chat and form relationships on their own - a deployable, hackable version of the 'g
An all-in-one, self-hostable desktop/Docker app for chatting with your own documents (RAG) and running configurable AI agents with tool access, aimed at running
An open-source LLM observability and evaluation platform - traces agent and RAG pipeline runs, runs automated evals on outputs, and surfaces where a pipeline is
An open-source AI data-analytics agent system built with DSPy - a team of specialized agents (data prep, statistics, visualization) that collaborate to answer a
A project-structure-aware autonomous agent for fixing bugs and resolving GitHub issues - it navigates the codebase using AST and code-search tools to localize t
One of the first widely-shared autonomous agent loops: given an objective, it generates, prioritizes and executes its own task list in a simple script rather th
IBM Research's open-source framework (Python and TypeScript) for building production-grade AI agents and multi-agent systems, with an emphasis on interoperabili
The open-source, self-hostable fork of StackBlitz's Bolt.new - generates a full web app from a prompt and runs it in an in-browser dev environment, but lets you
A research-oriented multi-agent framework for studying how communicative agents cooperate, negotiate and solve tasks together at scale, used as the base layer u
A customizable multi-agent framework that simulates a virtual software company - agents playing CEO, CTO, programmer and tester roles that converse through phas
A CLI coding assistant built specifically around Claude's tool-use API - it can create/read/edit files and run commands autonomously on a coding task, as a comm
A VS Code extension that runs an autonomous coding agent directly in the editor - it can create/edit files, run terminal commands, and use the browser, asking f
A React framework for adding AI copilots and agent-driven UI into your own application - chat sidebars, in-app actions the agent can trigger, and shared state b
An open-source, secure sandbox infrastructure for running AI-generated code - spins up isolated dev environments in milliseconds, aimed at agent builders who ne
An AI-native data-app development framework combining multi-agent workflows, natural-language-to-SQL, and RAG over your own data sources, aimed at teams buildin
ByteDance's open-source Deep Exploration and Efficient Research framework - a community-driven workflow for research agents that plan, search, and write up find
An open-source agentic AI software engineer that breaks a high-level instruction into steps, researches relevant information, and writes code to accomplish it,
An open-source LLM application development platform combining a visual workflow builder, RAG pipeline, agent capabilities and prompt IDE, aimed at teams shippin
Stanford's framework for programming (rather than hand-prompting) language models: you write modular Python code for a pipeline or agent and DSPy compiles/optim
An open-source runtime that gives AI agents secure, isolated cloud sandboxes to execute code in, so an agent can run and test code it writes without touching yo
An open-source framework/operating layer for building autonomous agents with persistent personality and memory, widely adopted in the crypto/Web3 space for buil
An open-source framework for financial large language models - data pipelines and fine-tuning recipes for tasks like sentiment analysis and forecasting on finan
An open-source deep-reinforcement-learning framework for quantitative finance - it provides market environments and training pipelines so researchers can train
A web-scraping API/engine built for AI use: point it at a URL or a whole site and it returns clean, LLM-ready markdown/structured data, handling crawling, rende
A free, open-source crypto trading bot you configure with your own strategy and run against an exchange - backtesting, hyperparameter optimization and dry-run m
Google's open-source terminal AI agent built on Gemini - reads and edits files, runs shell commands, and can call out to MCP tools, aimed at developers who pref
The original research code behind the 'Generative Agents' paper: a simulated town of AI residents with memory, reflection and planning, whose believable social
Google's open-source Agent Development Kit for building, evaluating and deploying multi-agent systems, designed to feel like standard software development (test
An AI pair-programmer agent (from Pythagora) that builds an app step by step, writing code, running and debugging it, and checking in with the developer at key
An autonomous research agent that plans a research question, gathers and cross-checks information from many web sources, and writes a cited report, aimed at red
An open-source crypto trading bot for running market-making and other automated strategies across many centralized and decentralized exchanges, aimed at quant t
An open-source general AI agent for complex, multi-step real-world tasks - research, coding, and computer/browser use - from Intelligent Internet, benchmarked a
A platform and SDK for building stateful AI agents and multi-step workflows that persist memory and can run long tasks in the background, aimed at agents that n
A self-hostable personal AI app that indexes your own notes, documents and the web to answer questions, and can run scheduled autonomous research tasks, positio
An open-source observability and evaluation platform for AI agents built around fast, structured tracing of every LLM call and tool step, aimed at teams that ne
A visual, drag-and-drop builder for LangChain-style flows and agents - wire together prompts, models, tools and retrievers on a canvas, then export the flow as
LangChain's low-level framework for building stateful, controllable agents as graphs of nodes and edges, giving developers explicit control over branching, loop
A Python multi-agent framework organized around lightweight 'Agent' objects that message each other and use tools, designed to be simpler to reason about than h
An open-source Large Action Model framework for turning natural-language instructions into automated web actions, aimed at building reliable web agents for QA t
LiveKit's SDK for building realtime, multimodal voice AI agents on top of its WebRTC infrastructure - handles low-latency audio streaming and turn-detection so
An open-source app for chatting with your own documents using locally-run LLMs, with no data sent to any external API, aimed at users who want document Q&A with
A lightweight Python framework (from the Prefect team) for adding structured, type-safe AI functions and agents into an existing codebase, favoring plain functi
A TypeScript agent framework (from the team behind Gatsby) covering agents, workflows, RAG and evals in one package, aimed at JS/TS teams who want an agent stac
An open-source memory layer for AI agents and assistants: it extracts, stores and retrieves relevant facts across conversations so an agent can remember a user
A multi-agent framework that assigns different agents standard software-company roles (PM, architect, engineer, QA) and has them collaborate through a standard
Microsoft's open-source framework for building multi-agent AI applications - agents that can converse with each other and with humans, call tools, and be compos
A radically simplified version of SWE-agent - about 100 lines of code with no special tools beyond a shell - built to show how much of SWE-agent's performance s
LangChain's open-source implementation of a 'deep research' agent - it plans a multi-step research task, delegates sub-questions to parallel searches, and synth
Gives an LLM a local code interpreter it can actually run - it writes and executes Python, shell and JavaScript on your machine to complete a task, functioning
An open-source generative process-automation framework: it records your screen, keyboard and mouse activity doing a desktop task, then uses that to train an age
OpenAI's lightweight, production-oriented SDK for building multi-agent workflows - agents, handoffs between them, guardrails and tracing - as the successor to O
An open-source, terminal-based AI coding agent (from the SST team) built to work with any model provider, with a focus on a fast, keyboard-driven TUI rather tha
An open-source platform for AI software-development agents that can modify code, run commands, browse the web and call APIs - essentially a full development env
An open-source general-purpose AI agent built to replicate the capabilities of closed agent products like Manus - planning and executing multi-step tasks with b
CAMEL-AI's optimized multi-agent workforce for real-world tasks - browsing, coding, document and data handling - built on the CAMEL framework and benchmarked on
A Python library that adds a conversational layer on top of pandas - ask a plain-language question about a dataframe and it writes and runs the pandas code to a
A high-accuracy RAG agent purpose-built for scientific literature - it searches, reads and cites specific passages across a set of papers to answer a research q
An LLM-powered penetration-testing agent that breaks a pentest into reasoning, generation and parsing modules, guiding (or automating) the step-by-step process
An open-source AI search engine built as an alternative to Perplexity - it understands a question, searches the web, and answers with cited sources, self-hostab
An open-source Python framework for building real-time voice and multimodal AI agents - it handles the audio/video pipeline plumbing (turn-taking, interruptions
A terminal-based AI coding agent built for large tasks spanning many files and long context - it plans the work up front, then executes with a diff-review step
A low-code, multi-agent orchestration framework that wraps AutoGen- and CrewAI-style patterns behind a simpler YAML/Python interface for defining agent teams an
A self-hostable app for asking questions about your own documents entirely offline/on-prem - no data leaves your machine - aimed at privacy-sensitive teams that
A Python agent framework from the Pydantic team that brings type-safe, validated structured outputs and dependency injection to agent building, aimed at develop
Microsoft's AI-oriented quantitative investment platform - data handling, model training and backtesting for the full quant-research workflow, from idea to a ev
An open-source RAG engine built around deep document understanding - it parses complex layouts (tables, PDFs, slides) more accurately than naive chunking, aimed
An LLM agent that generates and keeps documentation in sync with a codebase automatically - summarizing modules and functions and updating docs as the code chan
A framework that lets multimodal models operate a computer the way a human would - taking screenshots, deciding where to click or what to type - aimed at giving
Microsoft's open-source SDK (C#, Python, Java) for integrating LLMs into applications - plugins, planners, memory and multi-agent orchestration - aimed at enter
An early 'junior developer' agent that takes a product spec and generates a whole small codebase file-by-file, popularizing the 'smol' philosophy of simple, hac
Hugging Face's minimalist agent framework built around agents that write and execute Python code as their action format, rather than JSON tool calls, aiming for
Browserbase's AI browser-automation framework built on Playwright: instead of brittle CSS selectors, you describe an action in plain language ('act', 'extract',
Stanford's research-writing system that generates a Wikipedia-style article on a topic from scratch - it simulates multi-perspective interviews with sources to
AWS's model-driven agent SDK: instead of hand-coding a decision loop, you give the agent a model, tools and a prompt and let the model itself drive planning and
An open-source, self-hostable generalist AI agent that browses the web, writes code, and manages files and workflows on your behalf, positioned as an open alter
An open-source platform for building, deploying and monitoring AI agents through an API, aimed at teams who want to ship an agent as a product feature without b
An enterprise-oriented multi-agent orchestration framework for production deployments - structured workflows, swarms of cooperating agents, and integrations mea
Princeton NLP's agent that turns an LLM into a autonomous software engineer for fixing real GitHub issues - it gets a custom agent-computer interface for browsi
An AI agent that turns a GitHub issue into a pull request automatically - it plans the change, edits the relevant files, and opens a PR for review, aimed at sma
A multi-agent LLM framework that simulates a trading firm - separate agents for fundamentals, sentiment, technicals and risk debate a trade before a final decis
Fetch.ai's lightweight Python library for creating autonomous economic agents that can discover and transact with each other over a decentralized network, aimed
A Python framework for building reliable AI agents with built-in verification/guardrail layers and MCP tool support, aimed at production use cases where a raw a
LandingAI's agentic library for computer vision tasks: describe a vision problem in plain language and it plans and writes the vision pipeline code (detection,
An open-source library for building voice-based LLM agents - phone calls, web calls, and voice assistants - by composing your choice of speech-to-text, LLM, and
An open-source TypeScript framework for building AI agents with built-in observability - a visual console for inspecting an agent's reasoning, tool calls and su
A research agent that plays Minecraft continuously, writing and storing its own skill library of reusable code as it explores - an early, widely-cited demo of a
An open-source autonomous LLM agent designed to solve complex goals end-to-end, splitting work between a high-level dispatcher agent and specialized sub-agents
A long-term memory store built specifically for AI agents - it builds a temporal knowledge graph from conversation history so an agent can query facts about a u