Agent Lightning

Framework for LLM applications

The most widely-used framework for building LLM-powered applications, available in Python and JavaScript. Provides abstractions for chains, agents, RAG, memory, tool usage, and structured output. Integrates with 100+ LLM providers, vector stores, document loaders, and tools. LangSmith offers tracing and evaluation. LangGraph enables stateful, multi-agent workflows with cycles. 100K+ GitHub stars. The de facto standard for LLM application development despite growing alternatives like LlamaIndex.

Microsoft's conversational multi-agent framework

AutoGen is an open-source programming framework from Microsoft Research for building AI agents and facilitating cooperation among multiple agents to solve complex tasks through multi-turn conversations. Pioneered conversable agents that interact, use tools, and involve humans in the loop for multi-agent workflows. v0.4 features a redesigned async event-driven architecture with stronger observability, flexible collaboration patterns, and reusable components.

Multi-agent AI framework

Python framework for orchestrating autonomous AI agents that collaborate to accomplish complex tasks. Define agents with specific roles, goals, and backstories, then organize them into crews with sequential or parallel task execution. Supports tool usage (web search, file I/O, API calls), memory, delegation between agents, and human-in-the-loop input. Works with OpenAI, Anthropic, local models, and more. 25K+ GitHub stars. Leading multi-agent framework alongside LangGraph and AutoGen.

Stateful agent orchestration framework by LangChain

LangGraph is LangChain's framework for building stateful, multi-actor AI agent applications as controllable graphs. It models workflows as nodes and edges, enabling cycles, branching, and human-in-the-loop patterns that simple chains cannot express. Features built-in persistence for conversation memory, streaming support, and fault tolerance. Provides fine-grained control over execution flow while supporting single-agent and multi-agent architectures with shared or independent state.

Official Python SDK for OpenAI agents

OpenAI's Python framework for building multi-agent AI applications with GPT models. Provides primitives for creating agents with tool calling, handoffs between specialized agents, guardrails for input/output validation, and tracing for observability. Supports building complex workflows where agents collaborate on tasks. Includes built-in tools for file search, code execution, and web browsing. Designed for production agent systems with structured output and error recovery patterns.

Official MCP server for GitHub repo operations

GitHub MCP Server is the official Model Context Protocol server from GitHub that connects AI assistants to repositories, issues, pull requests, workflows, and code search. It exposes 100+ operations with toolset filtering, permission scoping, and audit logging, available in both remote-hosted and self-hosted Docker deployment modes.

Up-to-date docs for AI code editors via MCP

Context7 is an MCP server developed by Upstash that injects up-to-date, version-specific documentation directly into AI code editors and coding assistants. By typing 'use context7' in prompts, developers get accurate library documentation instead of hallucinated or outdated API references. It pulls from official source documentation and serves it through the Model Context Protocol, solving the common problem of LLMs generating code with incorrect or nonexistent API calls.

Apache-2.0 Python framework for repeatable LLM, RAG, agent, MCP, and safety evaluation workflows.

DeepEval is an Apache-2.0 Python framework for evaluating LLM apps, RAG systems, agents, MCP workflows, and safety behavior with repeatable test cases. It works locally and in CI/CD, then connects to Confident AI for hosted reports, observability, red teaming, and governance when teams need shared evidence instead of ad-hoc prompt reviews and manual QA.

Turn websites into LLM-ready structured data

Firecrawl is a Y Combinator-backed API that crawls websites and converts them into clean, LLM-ready Markdown or structured JSON. Handles JavaScript rendering, pagination, sitemaps, and anti-bot measures automatically. Designed for RAG pipelines, AI agents, and data extraction workflows. Features batch crawling, scheduled scraping, webhook notifications, and custom extraction schemas. Processes content for direct ingestion into vector databases and LLM context windows.

Virtual engineering team as Claude Code skills by YC CEO Garry Tan

GStack transforms Claude Code into a structured virtual engineering team through 23 opinionated slash command skills created by Y Combinator CEO Garry Tan. It assigns specialist roles including CEO product review, engineering manager architecture oversight, designer visual audit, QA lead with real browser testing, and release engineer deployment. Each skill enforces focused workflows with clear decision principles for running parallel coding sessions.

Zig-built headless browser engineered for AI agent workloads

Open-source headless browser written in Zig for AI agents, crawling, and automation. Lightpanda omits graphical rendering, keeps DOM and JavaScript execution, exposes CDP for Puppeteer/Playwright/chromedp, and adds Agent, PandaScript, and MCP workflows. Current public benchmarks claim about 9x faster execution and 16x less memory than Chrome.