Candit is the AI notetaker built for recruiters who are tired of getting fooled by candidates using live AI copilots, and cant decide on which notetaker to use. Real-time notes, ATS + MCP ready and built-in fraud detection.
I built Candit because AI has quietly changed technical interviews. A lot of candidates are now running live AI copilots on the call. It’s not just unfair to recruiters - it’s brutal for the candidates who actually prepared without one.
I spent weeks stalking recruiter-focused conversations across the internet, and the two biggest issues they were all facing: deciding on what notetaker to use and cheaters.
Traditional interview fraud detection is expensive, complicated, and usually forces candidates into locked-down, restricted environments. Meanwhile, a single bad hire can cost a company tens of thousands in wasted time, salary, and re-hiring.
Candit is different. You just bring it with you to any interview - no special setup, no restricted environment required. It works in the same Zoom, Google Meet, or Teams calls you’re already running.
Most AI notetakers only work after the call ends. You wait, then get notes. Candit writes structured notes live, while the conversation is still happening.
While it does that, it’s also watching for signs of an AI copilot in the room 👻 We flag it in real time - not in a report the next day.
How the detection works (briefly): We run localised, reverse-engineered versions of the most common stealth AI copilots in real time. We feed the live transcript into them and simulate what those tools would generate, then match patterns against what the candidate is actually saying. The models get updated weekly to stay accurate.
You can also drop your own notes mid-call and Candit folds them into the live summary (Granola-style), run </commands> to score a candidate or ask questions without leaving the meeting, connect it to your ATS to score against the actual role, and it speaks MCP if you want to wire it into anything else.
Everything happens in real time, so your workflow no longer has to wait for the meeting to end.
Would genuinely love to know: if your AI notetaker could do one more thing today, what would it be?
(Ad’s a bit rough - my motion designer failed me so I had to make it myself overnight! )
Solo founder here 👋 second Product Hunt launch.
I built Candit because AI has quietly changed technical interviews. A lot of candidates are now running live AI copilots on the call. It’s not just unfair to recruiters - it’s brutal for the candidates who actually prepared without one.
I spent weeks stalking recruiter-focused conversations across the internet, and the two biggest issues they were all facing: deciding on what notetaker to use and cheaters.
Traditional interview fraud detection is expensive, complicated, and usually forces candidates into locked-down, restricted environments. Meanwhile, a single bad hire can cost a company tens of thousands in wasted time, salary, and re-hiring.
Candit is different. You just bring it with you to any interview - no special setup, no restricted environment required. It works in the same Zoom, Google Meet, or Teams calls you’re already running.
Most AI notetakers only work after the call ends. You wait, then get notes. Candit writes structured notes live, while the conversation is still happening.
While it does that, it’s also watching for signs of an AI copilot in the room 👻 We flag it in real time - not in a report the next day.
How the detection works (briefly):
We run localised, reverse-engineered versions of the most common stealth AI copilots in real time. We feed the live transcript into them and simulate what those tools would generate, then match patterns against what the candidate is actually saying. The models get updated weekly to stay accurate.
You can also drop your own notes mid-call and Candit folds them into the live summary (Granola-style), run </commands> to score a candidate or ask questions without leaving the meeting, connect it to your ATS to score against the actual role, and it speaks MCP if you want to wire it into anything else.
Everything happens in real time, so your workflow no longer has to wait for the meeting to end.
Would genuinely love to know: if your AI notetaker could do one more thing today, what would it be?
(Ad’s a bit rough - my motion designer failed me so I had to make it myself overnight! )