AI agent frameworks compared

1. Rank 13

   Mastra

   Framework

   Builds TypeScript agents and graph-based workflows with memory, RAG, evals and model routing

   Pick it if you want agents, workflows, memory and evals in one TypeScript framework.

   • score 48
   • 28.8k stars
   • 36.1 stars/day since creation
   • pushed Oct 15
   • @mastra/[email protected] on Oct 14
   • Apache-2.0

   Create a project with npm

   npm create mastra@latest

   Copy install command for Mastra
   Use it when, skip it if, stats for Mastra

   Why you'd want it

   Mastra is a TypeScript framework for agents and workflows that plugs into React, Next.js and Node apps or runs as a standalone server. It connects to 40+ model providers through one interface, has a graph-based workflow engine with .then(), .branch() and .parallel(), and can suspend an agent or workflow for human input and resume from stored state. It also covers conversation history, RAG, observational memory, MCP server authoring, evals and observability.

   Best for

   • TypeScript and Next.js teams
   • Agents with human approval steps
   • Apps that need memory and RAG in one stack

   Use it when

   You want agents, workflows, memory and evals in one TypeScript framework instead of assembling separate libraries.

   Skip it if

   You work in Python. Features in ee/ directories (for example auth) need a Mastra enterprise license for production use.

   Stats

   Stars

   28,827

   Last push

   Oct 15

   Release

   @mastra/[email protected], Oct 14

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   36.1

   New and rising score

   48 of 100

   Full stat block for Mastra

2. Google Agent Development Kit (ADK)

   Framework Standard

   Builds, evaluates and deploys Python agents and graph-based workflows, optimized for Gemini

   Pick it if you want agent definitions, a local testing UI, evaluation and deployment to Cloud Run in one toolkit.

   • 21.9k stars
   • 38.9 stars/day since creation
   • pushed Oct 15
   • v2.12.0 on Oct 13
   • Apache-2.0

   Install with pip

   pip install google-adk

   Copy install command for Google Agent Development Kit (ADK)
   Use it when, skip it if, stats for Google Agent Development Kit (ADK)

   Why you'd want it

   ADK is a code-first framework where you define agents, tools and orchestration in Python, then compose them into multi-agent hierarchies or graph-based workflows with routing, fan-out/fan-in, loops, retries and human-in-the-loop. It ships a CLI with an interactive runner, a local web UI for testing and debugging, an adk eval command, and deploy commands for Docker and Cloud Run, with Vertex AI Agent Engine for scaling. It is optimized for Gemini but model-agnostic, and tools can come from custom functions, OpenAPI specs or MCP.

   Best for

   • Teams on Gemini and Google Cloud
   • Multi-agent hierarchies
   • Agents that need built-in evaluation

   Use it when

   You want agent definitions, a local testing UI, evaluation and deployment to Cloud Run in one toolkit.

   Skip it if

   You are on Python older than 3.11. If Python is not your stack, ADK has separate Java, Kotlin, Go and TypeScript versions in other repos.

   Stats

   Stars

   21,854

   Last push

   Oct 15

   Release

   v2.12.0, Oct 13

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   38.9

   Tier

   Standard, not ranked

   Full stat block for Google Agent Development Kit (ADK)

3. LangGraph

   Framework Standard

   Orchestrates long-running, stateful agents as graphs with durable execution and human-in-the-loop

   Pick it if you want explicit control over control flow and state, with checkpoints to resume from after a failure or a human review.

   • 43.3k stars
   • 37.2 stars/day since creation
   • pushed Oct 15
   • prebuilt==1.1.1 on Oct 15
   • MIT

   Install with pip

   pip install -U langgraph

   Copy install command for LangGraph
   Use it when, skip it if, stats for LangGraph

   Why you'd want it

   LangGraph is a low-level orchestration layer for agents and workflows that need to persist through failures and resume exactly where they left off. It provides short-term working memory and long-term memory across sessions, and lets a human inspect and modify agent state at any point during execution. It is built by LangChain Inc. but can be used without LangChain, and an equivalent JS/TS library, LangGraph.js, is available.

   Best for

   • Long-running stateful agents
   • Workflows with human review steps
   • Teams already using LangChain or LangSmith

   Use it when

   You want explicit control over an agent's control flow and state, with checkpoints it can resume from after a failure or a human review.

   Skip it if

   You want a quick agent with planning, subagents and file system use out of the box; the README points to Deep Agents, a higher-level package built on LangGraph, for that.

   Stats

   Stars

   43,287

   Last push

   Oct 15

   Release

   prebuilt==1.1.1, Oct 15

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   37.2

   Tier

   Standard, not ranked

   Full stat block for LangGraph

4. Microsoft Agent Framework

   Framework Standard

   Builds AI agents and graph-based multi-agent workflows in Python or .NET

   Pick it if you need multi-agent workflows with checkpointing and human review in Python or C#, or you are moving off AutoGen or Semantic Kernel.

   • 14.1k stars
   • 26.4 stars/day since creation
   • pushed Oct 14
   • dotnet-1.25.0 on Oct 14
   • MIT

   Install with pip

   pip install agent-framework

   Copy install command for Microsoft Agent Framework
   Use it when, skip it if, stats for Microsoft Agent Framework

   Why you'd want it

   Microsoft Agent Framework provides agents, middleware and graph-based workflows with sequential, concurrent, handoff and group collaboration patterns, plus checkpointing, streaming, human-in-the-loop and time-travel. It has built-in OpenTelemetry tracing, declarative YAML agents and a DevUI for testing. Microsoft's docs describe it as the direct successor to AutoGen and Semantic Kernel, built by the same teams, and the README links migration guides from both.

   Best for

   • Teams migrating from AutoGen or Semantic Kernel
   • Multi-agent workflows in .NET or Python
   • Azure and Microsoft Foundry deployments

   Use it when

   You need orchestrated multi-agent workflows with checkpointing and human review, in Python or C#, possibly hosted on Microsoft Foundry.

   Skip it if

   You want a minimal single-agent loop with no Azure ties; the README quickstarts use Microsoft Foundry and Azure credentials, and other providers are covered in samples. The Go SDK is a separate repo in public preview.

   Stats

   Stars

   14,132

   Last push

   Oct 14

   Release

   dotnet-1.25.0, Oct 14

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   26.4

   Tier

   Standard, not ranked

   Full stat block for Microsoft Agent Framework

1. Agno

   Framework Standard

   Builds Python agents, teams and workflows and serves them as an API with the AgentOS runtime

   Pick it if you want to run your agents as a multi-user service on your own infrastructure, with data in your own database.

   • 42.7k stars
   • 26.3 stars/day since creation
   • pushed Oct 15
   • v3.1.3 on Oct 12
   • Apache-2.0

   Install with pip

   pip install -U agno openai

   Copy install command for Agno
   Use it when, skip it if, stats for Agno

   Why you'd want it

   Agno is a Python SDK with three primitives (agents, teams and workflows) plus a runtime, AgentOS, that serves them through a production API with 50+ endpoints over SSE and websockets. Sessions, memory, knowledge and traces are stored in your own database, and the runtime adds human approval, OpenTelemetry tracing, JWT-based RBAC with multi-tenant isolation, cron scheduling, and interfaces for Slack, Telegram, WhatsApp, Discord, AG-UI and A2A. Starter templates deploy it with Docker to Railway, AWS, GCP, Azure, Fly, Render, Modal or Helm.

   Best for

   • Self-hosted agent platforms
   • Serving agents to many users
   • Agents exposed in Slack or messaging apps

   Use it when

   You want to build agents and also run them as a multi-user service on your own infrastructure, with data kept in your own database.

   Skip it if

   You are not on Python. Agno sends a telemetry event per agent run by default (prompts, messages and outputs are not sent); set AGNO_TELEMETRY=false to turn it off.

   Stats

   Stars

   42,736

   Last push

   Oct 15

   Release

   v3.1.3, Oct 12

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   26.3

   Tier

   Standard, not ranked

   Full stat block for Agno

2. Claude Agent SDK

   Framework Standard

   Embeds Claude Code's agent loop, built-in tools and context management in Python or TypeScript apps

   Pick it if you want Claude Code's tools and permission model inside your own process. It runs Claude models only.

   • 8.3k stars
   • 16.9 stars/day since creation
   • pushed Oct 14
   • v0.2.167 on Oct 14
   • Python repo MIT

   Install with pip

   pip install claude-agent-sdk

   Copy install command for Claude Agent SDK
   Use it when, skip it if, stats for Claude Agent SDK

   Why you'd want it

   The Agent SDK is a library that runs the Claude Code binary, so your agent gets the same tools (read, write and edit files, run commands, search the web), agent loop and context management that power Claude Code. You can add custom tools as in-process MCP servers, run hooks at points in the agent lifecycle, spawn subagents, control which tools run automatically through permissions, and resume or fork sessions. It is available for Python (claude-agent-sdk) and TypeScript (@anthropic-ai/claude-agent-sdk).

   Best for

   • Agents that edit files and run commands
   • Teams already using Claude Code
   • Embedding a coding agent in your own app

   Use it when

   You want Claude Code's tool set and permission model inside a process you operate, instead of writing the tool loop yourself.

   Skip it if

   You want to use non-Claude models. It needs an Anthropic API key or Claude through Amazon Bedrock, Google Cloud or Microsoft Foundry, and third-party products may not offer claude.ai login unless approved by Anthropic.

   Stats

   Stars

   8,283

   Last push

   Oct 14

   Release

   v0.2.167, Oct 14

   License

   Python repo MIT

   Pricing

   Paid

   Stars per day, since repo creation

   16.9

   Tier

   Standard, not ranked

   Full stat block for Claude Agent SDK

3. CrewAI

   Framework Standard

   Orchestrates teams of role-based agents (Crews) and event-driven workflows (Flows) in Python

   Pick it if you want specialized agents with clear roles and tasks, wrapped in event-driven workflows in Python.

   • 59.7k stars
   • 55.1 stars/day since creation
   • pushed Oct 15
   • 1.15.28 on Oct 14
   • MIT

   Install the CLI with uv

   uv tool install crewai

   Copy install command for CrewAI
   Use it when, skip it if, stats for CrewAI

   Why you'd want it

   CrewAI is a standalone Python framework with its own primitives for agents, tasks, crews, flows, tools and orchestration. Crews let role-based agents collaborate and delegate tasks autonomously, while Flows give event-driven control with state management and conditional branching, and the two can be combined. The crewai CLI scaffolds a project where agents and tasks are defined in JSONC files and run with crewai run.

   Best for

   • Multi-agent teams with defined roles
   • Business process automation
   • Mixing autonomous agents with deterministic steps

   Use it when

   You want several specialized agents with clear roles and tasks, wrapped in deterministic, event-driven workflows in regular Python.

   Skip it if

   You are not on Python (it requires Python 3.10 to 3.13). It collects anonymous telemetry by default (disable with OTEL_SDK_DISABLED=true), and the managed control plane, CrewAI AMP, is a separate commercial product.

   Stats

   Stars

   59,704

   Last push

   Oct 15

   Release

   1.15.28, Oct 14

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   55.1

   Tier

   Standard, not ranked

   Full stat block for CrewAI

4. Deep Agents

   Framework Standard

   Gives you a ready-made agent harness with planning, files, subagents and context management

   Pick it if you want planning, subagents, files and context management on day one, with every piece overridable.

   • 30.3k stars
   • 68.2 stars/day since creation
   • pushed Oct 15
   • deepagents==0.7.25 on Oct 14
   • MIT

   Install with uv

   uv add deepagents

   Copy install command for Deep Agents
   Use it when, skip it if, stats for Deep Agents

   Why you'd want it

   Deep Agents is LangChain's opinionated agent harness on top of LangGraph, with subagents that get isolated context windows, a filesystem over local, sandboxed or remote backends, shell access, summarization of long threads and human approval of tool calls. It works with any model that supports tool calling, including local models through Ollama, vLLM or llama.cpp, and accepts your own tools or MCP servers. A JS/TS version lives in a separate repo, deepagentsjs.

   Best for

   • Long-horizon, multi-step agents
   • Teams already on LangChain or LangGraph
   • Agents that need subagents and file access

   Use it when

   You want planning, subagents, files and context management working on day one, and still want to override any piece.

   Skip it if

   You need a custom control flow; LangGraph underneath gives you that directly. Its security model trusts the LLM, so you must enforce limits at the tool or sandbox level.

   Stats

   Stars

   30,277

   Last push

   Oct 15

   Release

   deepagents==0.7.25, Oct 14

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   68.2

   Tier

   Standard, not ranked

   Full stat block for Deep Agents

5. OpenAI Agents SDK

   Framework Standard

   Builds multi-agent workflows with handoffs, guardrails, sessions and built-in tracing

   Pick it if you want a small set of primitives (agents, tools, handoffs, guardrails) with tracing built in.

   • 30k stars
   • 51.5 stars/day since creation
   • pushed Oct 15
   • v0.24.0 on Oct 15
   • MIT

   Install with pip

   pip install openai-agents

   Copy install command for OpenAI Agents SDK
   Use it when, skip it if, stats for OpenAI Agents SDK

   Why you'd want it

   The OpenAI Agents SDK is a lightweight Python framework where agents are LLMs configured with instructions, tools, guardrails and handoffs to other agents. It is provider-agnostic, supporting the OpenAI Responses and Chat Completions APIs plus 100+ other LLMs, and includes sessions for conversation history, human-in-the-loop, and built-in tracing of agent runs. It also covers sandbox agents that work in a container over long tasks, plus realtime and voice agents.

   Best for

   • Multi-agent apps with handoffs
   • Voice and realtime agents
   • Teams on the OpenAI API

   Use it when

   You want a small set of primitives (agents, tools, handoffs, guardrails) with tracing built in, and may want voice or sandboxed agents later.

   Skip it if

   You work in JavaScript or TypeScript; use the separate Agents SDK JS/TS (openai/openai-agents-js) instead. The Python package needs Python 3.10 or newer.

   Stats

   Stars

   30,042

   Last push

   Oct 15

   Release

   v0.24.0, Oct 15

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   51.5

   Tier

   Standard, not ranked

   Full stat block for OpenAI Agents SDK

1. Pydantic AI

   Framework Standard

   Runs a typed Python agent loop with validated tools and outputs, and models swapped by a string

   Pick it if you want type checking and validation across tools and outputs, and to switch providers without rewriting the agent.

   • 20.6k stars
   • 24.4 stars/day since creation
   • pushed Oct 15
   • v2.56.0 on Oct 13
   • MIT

   Install with uv

   uv add pydantic-ai

   Copy install command for Pydantic AI
   Use it when, skip it if, stats for Pydantic AI

   Why you'd want it

   Pydantic AI gives you an agent whose tools and outputs are typed: tool arguments are validated before your code runs, and a run is guaranteed to return the output type you declared. The same agent can run behind a web frontend, in the terminal, on a voice call, or under durable execution with Temporal, DBOS or Prefect. Pydantic AI Harness adds capabilities such as memory, guardrails, sub-agents, planning and context management, and its instrumentation is plain OpenTelemetry.

   Best for

   • Typed data extraction
   • Python teams already using Pydantic
   • Durable background agents on Temporal or DBOS

   Use it when

   You want type checking and validation across tools and outputs, plus the option to switch model providers without rewriting the agent.

   Skip it if

   You are not working in Python.

   Stats

   Stars

   20,622

   Last push

   Oct 15

   Release

   v2.56.0, Oct 13

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   24.4

   Tier

   Standard, not ranked

   Full stat block for Pydantic AI

2. smolagents

   Framework Standard

   Runs agents that write their actions as Python code, with core agent logic in about 1,000 lines

   Pick it if you want a small, readable library whose agents act by writing code, and you can run that code in a sandbox.

   • 29.8k stars
   • 44 stars/day since creation
   • pushed Oct 6
   • v1.26.0 on May 29
   • Apache-2.0

   Install with pip

   pip install "smolagents[toolkit]"

   Copy install command for smolagents
   Use it when, skip it if, stats for smolagents

   Why you'd want it

   smolagents keeps abstractions minimal: the agent logic fits in roughly 1,000 lines, and its CodeAgent writes its actions in code. It is model-agnostic (local transformers or Ollama models, Hugging Face inference providers, or OpenAI, Anthropic and others via LiteLLM) and can use tools from MCP servers, LangChain or Hub Spaces. Agents and tools can be pushed to and pulled from the Hugging Face Hub, and a smolagent CLI runs agents from the terminal.

   Best for

   • Agents that act by writing code
   • Hugging Face Hub users
   • Reading and learning a compact agent loop

   Use it when

   You want a small, readable agent library and agents that act by writing code, with your choice of local or hosted models.

   Skip it if

   You cannot run agent code in a sandbox: the README says the built-in LocalPythonExecutor is not a security boundary and recommends E2B, Blaxel, Modal or Docker.

   Stats

   Stars

   29,849

   Last push

   Oct 6

   Release

   v1.26.0, May 29

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   44

   Tier

   Standard, not ranked

   Full stat block for smolagents

3. Strands Agents

   Framework Standard

   Runs a model-driven agent loop in your own process, in Python or TypeScript, with any model provider

   Pick it if you want to own the agent loop in your process, with limits, hooks and tracing built in.

   • 8.8k stars
   • 17 stars/day since creation
   • pushed Oct 14
   • python/v1.61.0 on Oct 14
   • Apache-2.0

   Install with pip

   pip install strands-agents strands-agents-tools

   Copy install command for Strands Agents
   Use it when, skip it if, stats for Strands Agents

   Why you'd want it

   Strands is an SDK for cases where you would otherwise write your own agent loop: it runs in your process with no hosted control plane and covers turn limits, token budgets, cancellation, tools, structured output, MCP, multi-agent patterns, memory, sessions, streaming, guardrails, tracing and evals. It has first-class support for Amazon Bedrock, Anthropic, OpenAI and Gemini plus other and custom providers, and hooks let you intercept any step of the loop. Strands harness (create_harness() in Python, createHarness() in TypeScript) gives a fully assembled agent with preset defaults.

   Best for

   • Replacing a hand-rolled agent loop
   • Amazon Bedrock users
   • Mixed Python and TypeScript stacks

   Use it when

   You want to own the agent loop in your own process, with lifecycle limits, hooks and tracing built in and the option to swap model providers.

   Skip it if

   You want a hosted agent platform rather than a library. The Python SDK needs Python 3.10+ and the TypeScript SDK needs Node.js 22+.

   Stats

   Stars

   8,817

   Last push

   Oct 14

   Release

   python/v1.61.0, Oct 14

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   17

   Tier

   Standard, not ranked

   Full stat block for Strands Agents

1. LangChain

   Framework Standard

   Provides one interface and a large integration library for models, embeddings, vector stores and tools

   Pick it if you want ready-made integrations for models, vector stores and retrievers behind one interface.

   • 148k stars
   • 101 stars/day since creation
   • pushed Oct 14
   • langchain-core==1.6.10 on Oct 12
   • MIT

   Install with pip

   pip install -U langchain

   Copy install command for LangChain
   Use it when, skip it if, stats for LangChain

   Why you'd want it

   LangChain is a framework for building agents and LLM apps from interoperable components, with a large library of integrations for model providers, tools, vector stores and retrievers. Its standard model interface lets you swap models as you experiment without rebuilding the app. An equivalent JS/TS library, LangChain.js, is available, and it pairs with LangGraph for lower-level orchestration and LangSmith for evals and tracing.

   Best for

   • Apps that need many model or data integrations
   • Switching between model providers
   • Prototyping LLM apps in Python or JS

   Use it when

   You want ready-made integrations for models, vector stores and retrievers, and a single interface for switching providers.

   Skip it if

   You need fine-grained control over agent flow and state; the README points to LangGraph for advanced customization and agent orchestration.

   Stats

   Stars

   147,843

   Last push

   Oct 14

   Release

   langchain-core==1.6.10, Oct 12

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   101

   Tier

   Standard, not ranked

   Full stat block for LangChain

2. LlamaIndex

   Framework Standard

   Connects LLM agents to your data with connectors, indexes, query engines and event-driven workflows

   Pick it if your agent's main job is answering questions or taking actions over your own documents and data sources.

   • 52.5k stars
   • 36.4 stars/day since creation
   • pushed Oct 14
   • v0.14.25 on Sep 21
   • MIT

   Install with pip

   pip install llama-index

   Copy install command for LlamaIndex
   Use it when, skip it if, stats for LlamaIndex

   Why you'd want it

   The LlamaIndex framework is a Python toolkit for context augmentation: data connectors ingest APIs, PDFs, docs and SQL, indexes structure the data, and query and chat engines retrieve context for the model. Agents can call a RAG pipeline as one tool among many, and Workflows add event-driven, multi-step processes with branching, retries and human-in-the-loop review. You can start with the llama-index starter package or install llama-index-core plus only the integrations you need from over 300 packages.

   Best for

   • RAG over private documents
   • Agents that query your own data
   • Pairing with LlamaParse for hard documents

   Use it when

   Your agent's main job is answering questions or taking actions over your own documents and data sources.

   Skip it if

   You want a framework that is its maintainer's main focus: the README says the company's primary focus has shifted to LlamaParse, while the OSS framework stays available.

   Stats

   Stars

   52,501

   Last push

   Oct 14

   Release

   v0.14.25, Sep 21

   License

   MIT

   Pricing

   Open source

   Stars per day, since repo creation

   36.4

   Tier

   Standard, not ranked

   Full stat block for LlamaIndex

3. Vercel AI SDK

   Framework Standard

   Gives TypeScript apps one API for text, structured output, tool-calling agents and chat UI hooks

   Pick it if you are building an agent into a web app and want the model calls and the chat UI from one library.

   • 27.3k stars
   • 22 stars/day since creation
   • pushed Oct 15
   • @ai-sdk/[email protected] on Oct 15
   • Apache-2.0

   Install with npm

   npm install ai

   Copy install command for Vercel AI SDK
   Use it when, skip it if, stats for Vercel AI SDK

   Why you'd want it

   The AI SDK is a provider-agnostic TypeScript toolkit with a unified API across providers such as OpenAI, Anthropic and Google. generateText covers text and Zod-typed structured output, ToolLoopAgent runs tool-calling agents, and AI SDK UI provides framework-agnostic hooks for chat and generative UIs in Next.js, React, Svelte and Vue. By default it routes model strings through the Vercel AI Gateway, and you can install provider packages to connect to providers directly.

   Best for

   • Next.js and React apps
   • Streaming chat UIs
   • Switching model providers in TypeScript

   Use it when

   You are building an AI feature or agent into a web app and want the model calls and the chat UI from one library.

   Skip it if

   You are not on JavaScript or TypeScript; it needs Node.js 22+. If you do not want the default Vercel AI Gateway routing, plan to install provider packages such as @ai-sdk/openai.

   Stats

   Stars

   27,323

   Last push

   Oct 15

   Release

   @ai-sdk/[email protected], Oct 15

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   22

   Tier

   Standard, not ranked

   Full stat block for Vercel AI SDK

1. AutoGen

   Framework Standard

   Builds multi-agent AI apps that act autonomously or alongside humans; now in maintenance mode

   Only if you already run it: AutoGen is in maintenance mode, and Microsoft recommends Microsoft Agent Framework for new projects.

   • 61.4k stars
   • 53.2 stars/day since creation
   • pushed Apr 15
   • python-v0.7.5 on Sep 30, 2025
   • MIT for code

   Install with pip

   pip install -U "autogen-agentchat" "autogen-ext[openai]"

   Copy install command for AutoGen
   Use it when, skip it if, stats for AutoGen

   Why you'd want it

   AutoGen is Microsoft's multi-agent framework, layered into a Core API (message passing, event-driven agents, local and distributed runtimes), an AgentChat API for common patterns such as two-agent chat and group chat, and an Extensions API for model clients and code execution. AutoGen Studio adds a no-code GUI for prototyping multi-agent workflows, though the README says it is not meant to be a production-ready app. Existing users can keep using it, but it no longer receives new features.

   Best for

   • Maintaining existing AutoGen projects
   • Studying multi-agent orchestration patterns

   Use it when

   You already run AutoGen and need it to keep working while you plan a migration.

   Skip it if

   You are starting a new project: AutoGen is in maintenance mode and community managed, and Microsoft recommends its successor, Microsoft Agent Framework, which has a migration guide from AutoGen.

   Stats

   Stars

   61,411

   Last push

   Apr 15

   Release

   python-v0.7.5, Sep 30, 2025

   License

   MIT for code

   Pricing

   Open source

   Stars per day, since repo creation

   53.2

   Tier

   Standard, not ranked

   Full stat block for AutoGen

1. Rank 2

   Reef

   Framework

   Turns feedback on an agent's live requests into versioned updates to its model weights or its harness

   • score 74
   • 7.6k stars
   • 168 stars/day since creation
   • pushed Oct 15
   • v0.1.3 on Oct 14
   • Apache-2.0

   Install from PyPI with uv

   uv venv && source .venv/bin/activate
   uv pip install reef-infra
   python3 -c "import reef; print(reef.__version__)"

   Copy install command for Reef
   Use it when, skip it if, stats for Reef

   Why you'd want it

   Reef sits between an agent and its model as an OpenAI- and Anthropic-compatible inference endpoint, records each interaction and returns a receipt that a later report can score or comment on. Recipes turn eligible feedback into updates, either training model weights with Slime and SGLang on a GPU stack or refining the harness, meaning prompts, rules and skills, through a model API with no local GPUs. Accepted updates go into a version history and are served without restarting Reef. The built-in Reefine recipe refines a coding harness from plain-language asks, such as reproducing a bug with a failing test first.

   Best for

   • Agents that should learn from user feedback
   • Harness optimization without GPUs
   • Continual learning research

   Use it when

   You run an agent with a steady stream of tasks or feedback and want that signal to produce evaluated, versioned improvements instead of manual prompt edits.

   Skip it if

   You want a stable agent today: harness work needs a model endpoint, representative tasks and an evaluator, weight training needs a supported GPU stack, and it is at v0.1.3 with the repo created on 2026-08-31. The harness install pipes a script into bash and the pi adapter runs commands with no sandbox, so keep the install root outside the project the agent works in.

   Stats

   Stars

   7,602

   Last push

   Oct 15

   Release

   v0.1.3, Oct 14

   License

   Apache-2.0

   Pricing

   Open source

   Stars per day, since repo creation

   168

   New and rising score

   74 of 100

   Full stat block for Reef

2. Rank 4

   OpenBot

   Framework

   Runs self-hosted AI coworkers with their own browser, files and shell, each action policy-checked and logged

   • score 68
   • 6.5k stars
   • 110 stars/day since creation
   • pushed Oct 14
   • v0.2.0 on Oct 12
   • MIT

   Run the published Docker image

   docker run -p 3001:3001 --env-file .env \
     -e EMBEDDED_POSTGRES=on -v openbot-data:/var/lib/postgresql \
     ghcr.io/copilotkit/openbot:latest

   Copy install command for OpenBot
   Use it when, skip it if, stats for OpenBot

   Why you'd want it

   OpenBot is a template for an in-house agent platform: Docker Compose brings up the app, PostgreSQL and the Bots, and you supply the model key. Each Bot gets its own container with a Chromium profile, a workspace volume and a shell, and every browser, file, shell, MCP or component action passes through one gateway that evaluates CEL policy rules, writes an audit row and only then acts. Any agent that speaks the AG-UI protocol, built with LangGraph, Mastra, CrewAI, Pydantic AI, Google ADK or by hand, can join as a coworker with its own channel, and you can watch its screen and take over. It ships with General Assistant and Knowledge coworkers plus example roles to copy.

   Best for

   • Self-hosted agent coworkers for a team
   • Auditable browser and shell actions
   • Bringing AG-UI agents into one workspace

   Use it when

   You want to give agents real computers inside your own infrastructure but need allow and deny rules and an audit trail for what they do.

   Skip it if

   You want a finished product: the README calls it an alpha template to clone and change, with no hosted version. It needs Docker, Bun, a model key and a CopilotKit Intelligence project (a free plan exists, or you can self-host it), and Bots drive a browser with saved logins and run shell commands, so set up sign-in before anyone else can reach it.

   Stats

   Stars

   6,491

   Last push

   Oct 14

   Release

   v0.2.0, Oct 12

   License

   MIT

   Pricing

   Open source

   Works with

   Docker, Web

   Stars per day, since repo creation

   110

   New and rising score

   68 of 100

   Full stat block for OpenBot

More guides

• MCP servers for coding agents, picked by job
• Open-source coding agents compared
• Claude Code alternatives, by what you need instead
• Agent observability and eval tools
• Category: Agent frameworks