Observed arrival · 2026-09-21
Benchable: benchmark runs with statistical verdicts
A self-hostable benchmark store that compares latency, build time, bundle size, throughput, and other metrics, then reports whether changes are statistically meaningful.
- For
- Teams tracking performance regressions in CI
- Worth noticing
- Benjamini–Hochberg correction reduced 15 preliminary flags across 312 metrics to 2.
Field notes
Benchable accepts benchmark runs with branch, commit, environment, metrics, percentile samples, and optional span timing. The page says the first run mentioning a metric creates it and infers properties such as units and directionality. Its examples include an idempotent CI submission and a response carrying both a regression count and machine-readable verdicts. The project identifies itself as built with Next.js, Postgres, and Drizzle, and provides a self-hosting repository.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue