Description
Bencher is a continuous benchmarking service. It runs performance benchmarks on bare metal, because the site says shared CI runners add enough noise that two identical runs can return very different results. The aim is to catch performance regressions before they merge, rather than letting noisy alerts be ignored. You can run benchmarks locally or in CI, track results over time and get alerted when a benchmark moves. The pricing page lists a free plan for open source projects with public projects and a public benchmark series, a Pro plan for performance-critical projects, and an Enterprise plan with single sign-on and on-premise deployment. The project's source is on GitHub, and there is a Discord community for questions.…
Preview
Pricing
- Free$0/month
- Pro$100/month
- EnterpriseCustom
Highlights
- Continuous benchmarking in CI
- Bare metal runners to reduce noise
- Track benchmarks over time
- Free plan for open source projects
- Enterprise options including SSO
Ratings & Reviews
No ratings yet
Be the first to rate Bencher.
People also visit

benchmark-checklist
Vet a perf measurement (limiter, tuning, limits, errors, repeatability
Bencher FAQ
What is continuous benchmarking?
Running performance benchmarks on every change, so a slowdown is caught before the change is merged.
Why run on bare metal?
The site says shared CI runners add noise, so Bencher runs benchmarks on bare metal to get results you can trust.
Is it free?
The Free plan is for open source projects and includes public projects and a public benchmark series.
What does Pro include?
Private projects, priority support, unlimited concurrent jobs and on-demand runners at $1.00 per hour, according to the pricing page.
Is there an enterprise option?
Yes. Enterprise adds single sign-on, on-premise deployment and dedicated onboarding.
What are the best Bencher alternatives?
Popular alternatives to Bencher include CodSpeed, JSBench.me and Gobenchdata. See all Bencher alternatives









