Frameworks

Haystack

Open-source Python framework from deepset for building production RAG pipelines and agents, with a modular pipeline architecture that predates the current agent framework wave.

Semantic Kernel

Microsoft's open-source SDK for building AI agents and orchestrating LLM calls in C#, Python, and Java, designed for enterprise .NET and Java codebases.

DSPy

Stanford's framework for programming language models declaratively, replacing hand-tuned prompt strings with modules that get automatically optimized against your own data.

Vercel AI SDK

Open-source TypeScript toolkit for building AI-powered applications and agents, with unified APIs across model providers and first-class React/Next.js streaming support.

Mastra

Open-source TypeScript agent framework from the team behind Gatsby, with built-in workflows, RAG, memory, and evals for building production AI agents in a JS-native stack.

LangSmith

LangChain's observability and evaluation platform for LLM applications, tracing every step of a chain or agent run and scoring outputs against test datasets.

StackSpot AI

Multi-agent platform covering the software development lifecycle, letting teams orchestrate agents for code generation, review, and deployment.

OpenAI Agents SDK

Lightweight open-source framework from OpenAI for multi-agent workflows with handoffs, guardrails and built-in tracing.

Pydantic AI

Python agent framework from the Pydantic team that brings type-safe, validated structured outputs and dependency injection to LLM apps.

Agno

Python framework (formerly Phidata) for building multi-agent systems with memory, knowledge and tools, plus a runtime for serving them.

smolagents

Hugging Face's minimal agent library where agents write and run their actions as Python code, working with any LLM.

Google ADK

Google's open-source Agent Development Kit for building, evaluating and deploying multi-agent systems, optimized for Gemini but model-agnostic.

Claude Agent SDK

Anthropic's SDK for Python and TypeScript that exposes the Claude Code agent loop, tools, hooks and MCP support for building custom agents.

Parsel

Research framework that breaks a programming task into natural-language function descriptions, then has a code LLM generate and test implementations.

Vibe Compiler

Self-compiling JavaScript tool that processes markdown prompt stacks through an LLM to generate code, then uses that process to rewrite itself.

DevOpsGPT

Open-source multi-agent system that turns natural-language requirements into working software, pulling and pushing code through Git. Around 6,000 stars.

LangChain

Open-source framework for building LLM applications, with integrations for over 100 model providers and a standard agent loop around create_agent.

LangGraph

LangChain's graph-based runtime for stateful agents, modeling each step as a node with checkpointing at every super-step for time-travel debugging.

CrewAI

Standalone Python framework for multi-agent systems, with no LangChain dependency, where agents get a role, a goal, and tools, then collaborate.

AutoGen

Microsoft Research's early multi-agent framework for Python, now in maintenance mode since October 2025, limited to bug fixes from community contributors.

LlamaIndex

Open-source data framework for RAG and agents over enterprise data, with 150-plus connectors through LlamaHub for sources like Notion and SQL.