Audio deepfake detection challenge · Open to all

Velma, Can You Catch This?

Velma is Modulate's ELM-powered audio deepfake detection model - currently ranked #1 on the Hugging Face Speech Arena leaderboard . Built for real-world audio, Velma is designed to catch what other detection tools miss.

Put it to the test. Submit your most challenging deepfake or a clip your current detection system failed to flag.

Get 1,000 free API credits

Drop your deepfake

Send us your best AI-generated voice, a convincing clone, or the audio your current tool completely missed. MP3 only · 10MB max · Name the file your X username (e.g. VoiceNinja99.mp3)

Drop your .mp3 file here

or click to browse your files

Clip submitted!

Format

MP3 only

Max size

10MB

Per user

1 submission

Modulate · Velma ELM API

Build with the world's best deepfake detector

Same model powering this challenge. Integrate Velma into your own product, research, or pipeline. No credit card needed.

Get 1,000 free API credits No credit card · instant access