Twilio Changelog | Oct. 01, 2026
Batch Transcription Configuration is now Generally Available!
Tune transcription without touching your code: engine, model, language, and destination live in a reusable configuration, so changing any of them is one update that every subsequent transcription picks up. Results are delivered via webhook or stored downstream in Conversation Orchestrator for persisted storage and Operator analysis.
Batch Transcription Configurations transcribes recorded calls after they complete, using a reusable configuration that sets your transcription engine, speech model, language, and destination. Choose Deepgram (Nova-3, Nova-2), or a Twilio-managed engine across up to 16 languages with many more to come. These APIs allow you to leverage automatic language detection on Deepgram, and map each audio channel to a participant so agent and customer speech stay separate.
For more information, please refer to the docs: