ZMIME
AssemblyAI · Speech recognition

AssemblyAI-STT-3.7-Realtime-Agent

Voice agents. Uses 32 languages in one stream and suppresses background conversation.

Specifications

Released
Nov 18, 2026
Inputs
Audio
Weights
Weights unconfirmed
Task
Speech recognition
Published capabilities
Voice agents. Uses 32 languages in one stream and suppresses background conversation.
Access
Realtime Speech-to-Text API as assemblyai-stt-3-7-realtime-agent
Result
Streaming transcript tuned for agent turn-taking
Billing terms
$0.45 USD per hour of audio on AssemblyAI.
Sources
Publisher documentation · Additional documentation

Compare AssemblyAI-STT-3.7-Realtime-Agent

AssemblyAI-STT-3.7-Realtime-Agent vs AssemblyAI-STT-3.7-Realtime