Voice agents built on OpenAI's Realtime API or a chained speech pipeline pass caller audio through your own server or client first. That is where Resemble Detect fits: stream the same audio to Detect and use the result to decide whether the agent should continue, verify, or hand off.
Stream caller audio to Resemble Detect from the same server that connects to OpenAI.
Use the result to end the session, hand off to a human, or require extra verification.
Resemble Identity can enroll voices from confirmed fraud calls and flag them when they call again, whatever caller ID or name they use.
Check recordings with Detect's file API for investigations.
