Tested · Open Source · Updated May 2026

Reviews, comparisons
& builds for AI systems.

Hands-on reviews of AI agent frameworks, head-to-head comparisons of LLM coding tools, and build tutorials with working code. Plus a live leaderboard of fifty models.

04 — Learn

Choose a learning path.

A curriculum if you want one, a reference if you don't. Pages are short by design and cross-linked.

Beginner start here

Set up Python, understand what an agent actually is, write your first ReAct loop in under fifty lines.

Intermediate design

Recurring agent shapes, the constraints token budgets impose, and how to specify behavior in a version-controlled way.

Advanced production

Multi-agent coordination, the protocol layer, and the trade-offs that matter at scale across the major frameworks.

05 — About

A reference, not a course.

This is a developer-focused reference for building with large language models. It covers AI agents, agentic workflows, the Model Context Protocol, and the Python frameworks people actually reach for in 2026 — LangChain, CrewAI, and AutoGen. Every concept is paired with runnable code, every framework page explains when to pick it and when to skip it, and every architectural pattern is grounded in a specific failure mode it solves.

The goal is to give you a working mental model fast, then hand you the code to try it yourself. There is no signup, no paywall, and no gated chapters — the entire guide is open and indexable. Pages sit in the 1,000 to 1,700 word range and cross-link aggressively, so following a link rarely sends you down a long detour. APIs change, so check the framework's own docs if something looks off; the site is updated as the ecosystem moves.

If you are new to all of this, start at Getting Started — it sets up the prerequisites and walks through your first agent in under fifty lines of Python.