Find your next framework
32LangGraph
Build agents as graphs. Loops, memory, human approval and full control of the flow.
LangChain
The layer under most agents. Models, tools, prompts and messages in one interface.
Google ADK
Google's Agent Development Kit, built around Gemini and Vertex AI.
OpenAI Agents SDK
OpenAI's own agent loop, with handoffs, guardrails and tracing built in.
Claude Agent SDK
Claude Code as a library. File editing, bash and search, ready to run.
Semantic Kernel
Microsoft's SDK for wiring models into .NET and Python applications.
Vercel AI SDK
The TypeScript way to build streaming AI apps and agents on the web.
CrewAI
Give each agent a role, a goal and a backstory, then let the crew work the task.
AutoGen
Microsoft's framework where agents solve problems by talking to each other.
MetaGPT
Agents assigned software company roles: PM, architect, engineer, reviewer.
CAMEL-AI
Research-led framework for studying how large agent societies cooperate.
LlamaIndex
Agents that reason over your documents. Strong on retrieval and indexing.
Haystack
deepset's pipeline framework for search and retrieval heavy applications.
DSPy
Stanford's approach: stop writing prompts, compile and optimise them instead.
Pydantic AI
Type safe agents. If you already trust Pydantic for validation, this feels like home.
Instructor
Get structured, validated objects back from any model instead of parsing prose.
n8n
Drag nodes onto a canvas and wire an AI workflow without writing much code.
Langflow
A visual builder for LangChain flows. Drag, connect, export as code.
Flowise
Open source drag and drop builder for LLM apps and agents.
Dify
A full platform: build, deploy and monitor agents from a browser.
AutoGPT
The project that made autonomous agents famous. Give a goal, walk away.
Agno
A fast, lightweight framework for multimodal agents with memory built in.
Letta
Formerly MemGPT. Agents that manage their own long term memory.
Griptape
Structured agents, pipelines and workflows with control over every step.
smolagents
Hugging Face's tiny agent library. Agents that write and run Python.
Browser Use
Point an agent at a real browser and let it click, type and read pages.
MCP
Not a framework. The standard that lets any agent plug into any tool or data source.
Composio
Hundreds of ready made tool integrations your agent can call, with auth handled.
LiteLLM
One interface and one proxy in front of a hundred model providers.
E2B
Secure cloud sandboxes so an agent can run the code it just wrote.
AgentOps
Tracing, cost tracking and replay for agents you have already shipped.
Mastra
The TypeScript agent framework, with workflows, evals and deployment included.