AI Gateway now supports streaming transcription
Streaming transcription just went live. Low-latency voice input for agents—no more waiting for the full audio file.
Why it matters
Vercel's AI Gateway adds streaming transcription in beta, enabling real-time voice-to-text for live captioning and voice agents with sub-second latency. This lowers the friction for building conversational AI products and agent voice modes across OpenAI, Anthropic, and xAI models.
The key facts
7 to knowStreaming transcription now beta on AI Gateway
Works with OpenAI GPT-realtime-whisper, Anthropic, xAI Grok-STT, and other streaming-capable models
Enables live captioning and voice agent input with reduced latency
Available via AI SDK streamTranscribe function
Returns partial and final transcript deltas in real-time
Allows voice mode for text-based agents without agent code changes
Works across multiple providers with single model string swap
Go to the source
Vercel Blogvercel.com
Publisher excerpt: AI Gateway now supports . Previously, transcription required a complete audio file and returned the full transcript in a single response. Now you can stream audio in as it's captured and get transcript updates back as the model produces them, keeping latency low for uses like live captioning and…