Catch the AI
on your call.
Real-time AI detector for calls — right in your browser.
Runs in your browser · Audio never leaves your device
Powered byHireduce
Hireduce“It’s important to note that there are several key factors…”
Three steps to the truth
Press the button
The detector opens in a compact window on top of your call.
Share computer audio
Pick the meeting tab and enable “share tab audio” to capture the conversation.
Watch
The AI-probability score and dials update live.
Analyze audio & video files
Upload a recording — MP3, WAV, M4A, MP4, MOV, and more. We transcribe it and run the same AI-detection analysis. Video files are converted to audio right in your browser before upload.
MP3 · WAV · M4A · OGG · WebM · FLAC · MP4 · MOV
Try file analysisSign in with Pro to upload files on your Home dashboard.
interview_recording.mp4
What the detector does
Listens to computer audio
Captures the sound of a browser tab or your system — so you hear the other person, not just yourself.
Live transcription
Turns the conversation into text instantly with Whisper — Groq cloud or fully in your browser.
AI-pattern analysis
Three dials score AI phrasing, spontaneity and bookishness of the speech.
Clear verdict
One probability, with the exact signals that fired. The call stays yours.
Frequently asked questions
Does it listen to my microphone or the computer’s sound?+
In “Computer audio” mode it captures the sound of a browser tab/system (the other person on the call). In “Microphone” mode it transcribes your mic instead.
Where is the audio processed?+
Depends on the engine: Groq and Server send audio to a server, while Local runs Whisper entirely in your browser.
Which browser do I need?+
A Chromium-based browser (Chrome or Edge). Tab-audio capture and the speech model work best there.
Is the verdict proof that someone used AI?+
No. It’s a probabilistic helper based on language patterns. Treat it as a hint, not evidence.
What languages are supported?+
Over 20 — English, Spanish, French, German, Italian, Portuguese, Ukrainian, Polish, Turkish, Arabic, Chinese, Japanese and more. Pick one in the detector.
Is anything recorded or stored?+
No. Audio is processed on the fly and the transcript lives only in the open window — nothing is written to a database.
How does the score work?+
Three independent dials — AI phrasing, spontaneity and bookishness — combine into one probability. AI marker phrases push it up; natural fillers and disfluencies pull it down.
Simple pricing
Start free. Upgrade when you need more.
Pro
- Unlimited hours
- All detection features
- File upload analysis
- Priority support
Try it on your next call.
Opens in seconds, right in your browser.