Sarvam Ships Realtime Streaming STT API With Mid-Call Reconfig and Millisecond VAD Tuning
itsOmSarraf_ · x · 2026-09-10
Indian AI startup Sarvam rolled out a batch of model and API updates in Aug-Sep, headlined by a new Realtime Streaming Speech-to-Text endpoint:
- GET /speech-to-text-realtime/ws, powered by saaras:v3-realtime, with saaras:v4-realtime now GA on the same protocol.
- Unlike the legacy streaming API (final transcript per utterance only), it emits true interim partial transcripts as the user speaks — built for voice agents that need barge-in and live captioning.
- Supports mid-stream reconfiguration via config.update (no reconnect), and simplifies VAD tuning from 10+ frame-count params to 3 millisecond-based ones.
- Also adds manual turn control via endpointing="manual", per-partial auto language detection, and structured error events.
Docs ship an /llms.txt index for AI agent consumption.
More from Models
- GPT-6 'Astra' Does 34 Math Steps in Latent Space, 4x More Than Sol — MaartenBaert · 2026-09-10
- NVIDIA details Alpamayo 2 Super, its L4 autonomy model for robotaxis — drmapavone · 2026-09-10
- OpenAI users report usage quotas wiped to zero as weekly reset dates shift by two days — ___Patrice___ · 2026-09-10
- Prediction: V4.1-Flash to score 36-38 on new AA index, agency at 42 — teortaxesTex · 2026-09-10
- Perplexity benchmarks 13 retrieval models: pplx-embed-v1-4b leads two of three categories — perplexity_ai · 2026-09-10
- Meme Budget: 'GPT-8 Swarm' Eats 182M GPUs as Astra Weighs Pausing Pro Signups — burny_tech · 2026-09-10