Independent · Sourced · Neutral
The examiner’s bench for AI agents and MCP.
We test the tools agents actually run on — Model Context Protocol servers and agent platforms — and publish the capabilities, pricing, and trade-offs, every fact sourced, so you can evaluate before you build.
- 2,500
- MCP servers cataloged
- 19
- Platforms examined
- 19
- Head-to-head matchups
- 0–5
- One rubric, every platform
Two shelves, one bench
Directory
MCP Servers
Reviews and comparisons of Model Context Protocol servers — capabilities, transport, auth, and how they behave under real workloads.
- Star-ranked catalog
- Transport & auth facts
- Official vs community
Directory
Agent Platforms
Hands-on evaluations of AI-agent platforms — pricing, model support, orchestration, and the trade-offs that matter when you build.
- Sourced scorecards
- Head-to-head matchups
- Best-of rankings
Latest analysis
All news →- 013 July 2026The State of MCP Servers 2026The Model Context Protocol registry now lists 13,108 servers — but the ecosystem is a long tail with almost no middle. Analyzing the 2,500 most-starred servers, we found it is overwhelmingly local-first (83% run on your machine), radically decentralized (91% of publishers ship a single server), and JavaScript-first (npm outnumbers Python 2-to-1).
- 022 July 2026Who leads on scale: why the infrastructure platforms top the scalability rankingsFour platforms earn a perfect scalability score. Three of them are infrastructure. Here's what a 5/5 on scale actually reflects.
- 031 July 2026Welcome to The Agent ExaminerWe are building a hands-on review site for AI-agent platforms and MCP servers. Here is what to expect.
- 0430 June 2026Alfe launches managed agents with 50% off the first three monthsAlfe is rolling out managed agents with an introductory 50% discount on the first three months. Here are the terms, the pricing model behind them, and the usual caveats.