If you build voice apps on Twilio, you already control the call flow, which makes it the natural place to check whether a caller's voice is real. Media Streams forks the call audio to a WebSocket you run. From there you pass it to Resemble Detect's streaming API and get results back while the caller is still talking. Your app decides what to do with them: keep going, ask an extra verification question, or route the call to a person.
Fork inbound, outbound, or both tracks to your server and forward the audio to Detect without changing how the call is answered.
Detect returns a label and score for each few seconds of audio, so you can branch in TwiML or your backend as the call goes on.
Run Resemble Detect on inbound audio to identify AI-generated or cloned voices before they reach your agents.
Check the caller before your Twilio-based voice agent shares account details or takes an action.
Point Detect at Twilio recording URLs after the call when you want a second look or an audit trail.
All audio detection processing is compliant with healthcare and data privacy requirements. Suitable for regulated industries.
