API Response Benchmark
Hammer an endpoint with a controlled burst of requests and get the latency distribution that actually matters — p50, p95, p99 and success rate.
Self-imposed safe ceiling: 100 requests, sequential. This is a diagnostic, not a load tester — for load testing use the Load Test Generator.
📊 Percentiles
Optimise for p95/p99, not the average. If p50 is 40ms but p99 is 900ms, a slice of your users is having a bad time on every page.
🔐 CORS
The endpoint must allow cross-origin requests from your browser, or the browser blocks the response. Timing of blocked requests is still captured but success rate will read 0%.
Frequently Asked Questions
Why does success rate show 0% but I still get timings?
If the endpoint doesn't send CORS headers, the browser blocks reading the response (so it can't confirm success) but the round-trip is still timed. Add Access-Control-Allow-Origin on a test route to see true success rates.