Gradium TTS lands on Baseten: streaming speech synthesis with 50ms TTFA
baseten · x · 2026-10-14
Baseten announced that Gradium TTS is now in its Model Library, pairing ultra-low 50ms time-to-first-audio with natural voices. It can be combined with any transcription model and LLM to run a full voice-agent pipeline on Baseten.
Technical details: it's a proprietary streaming TTS — no fixed context or output-duration caps, suited to real-time and long-form speech. Commercial use of outputs is permitted under Gradium's ToS. The model runs over WebSocket: the client sends a JSON setup (voice, output format), streams text, and signals end-of-text, while the server streams back the sample rate followed by raw audio chunks. Docs include a Python example using GradiumClient, deployed on H100.
More from Infra
- Espresso now auto-optimizes Databricks Spark jobs and configs — mattturck · 2026-10-14
- Intel Diamond Rapids motherboard shown off: 16 DDR5 slots, MRDIMM at 12800MT/s — firstadopter · 2026-10-14
- SayGM claims 70% cheaper LLM gateway with $5M ARR and crypto referral rewards — markjeffrey · 2026-10-14
- Eaton's 800 VDC power rack draws big crowds at OCP summit — firstadopter · 2026-10-14
- Mixed-precision, mixed-accelerator samplers cut RL training cost with off-policy correction — xiuyu_l · 2026-10-14
- Pairing M5 Ultra 256GB with M3 Ultra 512GB for disaggregated prefill/decode cuts TTFT from 40s to 18s — Usual_Tackle5892 · 2026-10-14