Nothing to set up.

Most cheat-detection tools work by locking the candidate down — a dedicated testing portal, a lockdown browser that blocks other tabs, a webcam-monitoring app they have to download and grant permissions to before the interview even starts. It's friction the candidate feels immediately, and IT overhead your team has to support.

It also solves the wrong problem. A stealth copilot doesn't need the browser tab the candidate is interviewing in — it runs alongside it, invisible to anything watching the screen.

Candit reads the conversation itself, not the candidate's screen — so there's nothing to lock down. You send a normal meeting link, on whatever video call tool you already use, and detection runs for the whole call without a single extra step from anyone on it.

1. 01
   Join meetings like you always do

   Paste a meeting link or connect your calendar — Candit joins on its own, same as any other participant. No plugin for you, no app for the candidate.

2. 02
   It announces itself by name

   Candit identifies itself when it joins and flags every session as recorded, same as any notetaker. That's the whole disclosure requirement.

3. 03
   Detection runs the whole time, quietly

   No separate step to turn it on. Every session on Plus, Pro, or Team gets AI Detectability by default.

4. 04
   Evidence lands where you already work

   Flags show up on the session the moment they happen, and push into Lever, Ashby, or any MCP-compatible tool.

Let candidates know a notetaker's in the meeting and listening — Candit already does that for you, announcing itself by name and flagging every session as recorded. That's the whole requirement; you don't need to separately disclose that AI Detectability is running.

Most agencies already screen this way — think of it less as surveillance, more as the safety net that keeps a bad hire from ever reaching a client.