Last updated 2026-10-27
Testwitness runs the tests you write, in plain words, against a website or a mobile app: an AI tester drives a real browser or a simulator, decides whether what it saw matches what your test says, and shows you the result with screenshots and a transcript. Mobile apps, and websites when you choose so, run on a Mac of yours that has the helper app installed.
You sign in with your email address. One account is one person. You are responsible for what is done with your account and for the projects, tests and credentials you put in it.
The AI tester uses a model from a provider you choose — Anthropic, OpenAI or Google Gemini — with an API key you bring. Every model call is made with your key and billed by that provider to you under its own terms. Testwitness never pays for model use on your behalf and never uses your key for anything but your own runs.
Only test websites and apps you own or are authorised to test. Test accounts you store are for that purpose alone. Don't use the service to attack, overload or scrape systems that aren't yours.
Testwitness is free to use for now, with a daily limit on how many runs an account can start. That period may end, after which starting runs will need a paid plan. When it ends, your projects, tests, notes and results stay available to read and edit; only starting new runs is affected. We will say so on the site before it happens.
The service is provided as is. An AI tester can be wrong: a verdict is a report of what it saw, not a promise about your software. The service may be unavailable or change at any time, and we may suspend an account that breaks these terms.
These terms may change. The date above says when they last did. Continuing to use the service after a change means you accept it.
Christian Arby
Rödhakevägen 6b
132 45 Saltsjö-Boo
Sweden