Coding tests were built to catch AI use. codesolara was built for a world where everyone already has it. Every candidate works in a real environment — VS Code in the browser, a terminal, and an AI pair called Solara — and you get a scorecard that shows how they actually think: their judgment, how they iterate, and whether they catch the assistant when it is wrong. Every score names the moment in the session it came from, and a second pass goes back and re-checks each one. No candidate account. No whiteboard. No pass/fail guesswork. Just a realistic task and an evidence-backed read your hiring committee can defend. Free for the whole team, unlimited seats — you pay when a candidate actually interviews. We're onboarding new teams now.