Checksum builds full end-to-end test coverage for a codebase by pairing two specialized AI agents: one generates Playwright and API tests from real user sessions and application behavior, and the other watches for breakage and auto-heals tests in real time as the underlying product changes. Every generated test runs as standard Playwright code committed to the team’s own repository, and a human engineer gives final verification before a test is considered ready, so a team is not left trusting AI-written tests unchecked. Founded in 2022 in San Francisco by Gal Vered, the company is bootstrapped without outside venture funding and states it helps engineering teams ship roughly ten times faster by removing manual test-writing and maintenance work.
Checksum sells itself entirely through a requested demo rather than a self-serve signup or trial, so a team cannot try it hands-on before talking to sales, and no plan-level feature breakdown is published alongside the price. The human-verification step is a meaningful part of the workflow’s reliability, worth asking about directly in a demo to understand how much of the final test suite is AI-written versus human-reviewed for a given codebase.
Pricing is Results-as-a-Service: no per-seat fees and no per-run charges, only a single price tied to the number of workflows or tests Checksum is actively maintaining for a team. Exact rates are not published and require a requested demo to get a quote.
