A measurement harness for how AI fetch pipelines handle content negotiation, compression, image formats, script execution, caching and timeouts.
Each test run gets its own URL namespace with unique canary tokens, so results from different agents never collide.
Then paste the run URL into the agent you want to test, asking it to visit and follow the instructions.
See what your own client looks like on the wire: https://ai-tester.paulcalvano.com/proto
Summary across every run: https://ai-tester.paulcalvano.com/reports