Self-service probing is free forever, and your key lives only in your local environment variables — never uploaded. There are only two things we charge for: ① you don't want to burn your own tokens / set up the environment, so we run the expensive jobs on our own infra (128K long context, cross-checks against the official API, model-identity fingerprinting); ② you want reports privately hosted, branded, re-tested on a schedule, with change alerts. Benchmark dimensions only — we never touch your business data.
What you pay for is us running the probes for you + hosting reports, not locking up the open-source capabilities — self-service probing and public reports are free forever and reproducible. Billing and payment channels are coming soon.
Want managed probing or Pro report hosting? Tell us now which gateways and which dimensions you want tested. We'll notify you as soon as payment and self-service ordering are available.
This toolkit and reporting platform does one thing — make gateway/model benchmarking solid and open. If it helped you pick the right gateway, come see the fuller capabilities in the main project.