The sampler platter of modern agent frameworks

What it does

This repository collects more than 80 practical examples, tutorials, and recipes for building LLM-powered applications. It bundles starter templates for text agents, voice assistants, RAG pipelines, and MCP-backed tools across a wide range of frameworks. The goal is to give developers working code they can run and extend rather than architecture diagrams to admire.

The interesting bit

The project treats framework fatigue as the problem to solve. Instead of evangelizing one stack, it places LangChain, CrewAI, PydanticAI, AWS Strands, Mastra, smolagents, and a dozen others side by side under the same roof. It even includes an eight-lesson AWS Strands course and sponsor-backed integrations, which is more curriculum than you usually get from an example collection.

Key highlights

• 19 starter agents spanning major and emerging frameworks (LangGraph, Letta, Semantic Kernel, Google ADK, KAOS, and others)
• Dedicated sections for voice agents, memory agents, RAG apps, MCP-backed tools, and browser automation
• An 8-lesson AWS Strands course covering fundamentals through production guardrails and multi-agent orchestration
• Simple-agent recipes for practical tasks like finance tracking, human-in-the-loop safety, and intelligent model routing
• Active Hacktoberfest participation and a sponsor-supported maintenance model

Verdict

Bookmark this if you are shopping for an agent framework or need a quick reference for wiring up RAG, voice, or MCP pipelines. Look elsewhere if you want a single, opinionated production framework rather than a broad survey.