Inspiration
AI agents are everywhere, but there's no way to verify they actually work. No standard testing. No directory of trusted agents. No reputation system. Google's A2A protocol gives agents a shared language — but a shared language doesn't guarantee trust.
What it does
A2Apex is a testing, certification, and directory platform for AI agents built on Google's A2A protocol.
Test — Point it at any A2A agent URL. We run 50+ automated compliance checks: Agent Card validation, live endpoint testing, state machine verification, streaming, auth, error handling.
Certify — Agents receive a 0-100 trust score and earn Gold, Silver, or Bronze certification with embeddable badges.
Get Listed — Every tested agent gets a public profile in the Agent Directory with trust scores, skills, test history, and badges.
How we built it
Python/FastAPI backend, vanilla JavaScript frontend, SQLite storage. No React, no build tools. Built in 2 weeks (15,000+ lines of code) using an AI coding assistant, running on a Mac mini in Wyoming.
Challenges we ran into
The A2A spec is brand new — there were no reference implementations to validate against. We had to build the compliance test suite from the raw spec document.
What we learned
The moat isn't in the code — it's in shipping fast and understanding the problem deeply. AI tools made the building possible; domain knowledge made it valuable.
What's next for A2Apex
- CI/CD integration for automated testing on every deploy
- Python SDK (pip install a2apex)
- Bulk testing for teams managing multiple agents
- Historical trust score tracking
