AI agents are reshaping how we build software. GitHub stars are a strong indicator of developer trust and community adoption. Here are the top 10 most starred AI agent frameworks heading into 2026.
Short on time? Get the key takeaways of this article in under 5 seconds with DocuMentor AI, my Chrome extension. No account required. Try it Here
1. LangChain ⭐ 122,850
langchain-ai/langchain | Python | MIT
The most popular framework for building LLM-powered applications, with extensive tooling for chains, agents, and retrieval.
2. MetaGPT ⭐ 61,919
FoundationAgents/MetaGPT | Python | MIT
A multi-agent framework that simulates a software company, with agents taking on roles like product manager, architect, and engineer.
3. AutoGen ⭐ 52,927
microsoft/autogen | Python | CC-BY-4.0
Microsoft's framework for building multi-agent conversational systems with customizable agent behaviors.
4. LlamaIndex ⭐ 46,100
run-llama/llama_index | Python | MIT
The leading framework for connecting LLMs to your data, with powerful indexing and retrieval capabilities.
5. CrewAI ⭐ 41,871
crewAIInc/crewAI | Python | MIT
Framework for orchestrating role-playing autonomous agents that collaborate to accomplish complex tasks.
6. Agno ⭐ 36,414
agno-agi/agno | Python | Apache-2.0
A multi-agent framework with a runtime and control plane for managing agent deployments at scale.
7. Haystack ⭐ 23,741
deepset-ai/haystack | Python | Apache-2.0
Production-ready AI orchestration framework focused on building customizable LLM applications and RAG pipelines.
8. Vercel AI SDK ⭐ 20,400
vercel/ai | TypeScript | Apache-2.0
The AI toolkit for TypeScript from the creators of Next.js, designed for building AI-powered web applications.
9. Open-AutoGLM ⭐ 19,800
zai-org/Open-AutoGLM | Python | Apache-2.0
An open-source phone agent model and framework for building mobile device automation agents.
10. Mastra ⭐ 19,021
mastra-ai/mastra | TypeScript | Apache-2.0
A TypeScript-native AI agent framework with built-in workflows, tools, and integrations.
Special Mentions
Stars don't tell the whole story. These frameworks have fewer stars but are gaining momentum with high download counts and modern architectures.
- OpenAI Agents SDK ⭐ 18,022 — Official OpenAI framework for multi-agent workflows
- Google ADK Python ⭐ 16,800 — Google's new agent development toolkit
- Pydantic AI ⭐ 14,000 — The Pythonic way to build AI agents
Key Takeaways
- Python dominates: 8 out of 10 top frameworks are Python-based
- LangChain leads by a massive margin with 2x the stars of the runner-up
- TypeScript options: Vercel AI SDK and Mastra for JS/TS developers
- Watch the special mentions: Momentum matters more than legacy stars
What to Read Next
- The Future of AI Building: Workflows, Agents, and Everything In Between
- LangGraph vs LlamaIndex: Who Makes AI Agents Easier in JavaScript?
- Getting Started with OpenAI's Agents SDK for TypeScript
Enjoying content like this? Sign up for Agent Briefings, where I share insights and news on building and scaling AI agents.
Star counts as of January 2026. Numbers change—check the repos for the latest.
Top comments (0)