

Tofu-1
#2 v Syntéza řeči (TTS)Bland Ai · v1 · od 2026-11-15 · 6× · naposledy 17. 11. 2026
92
Momentum
Tofu-1 is a full-duplex speech-to-speech model from Bland AI for phone calls with poor audio quality. It processes audio directly without an intermediate transcript and, according to the vendor, also generates background sounds and multiple speakers. It has been available since November 15, 2026 as a closed enterprise research preview; general availability is announced for January.
Vývoj momenta
23.08.21.11.
Vlastnosti
| Real-Time Streaming | Full duplex: listens and speaks simultaneously, caller can interrupt mid-sentence |
| Platform | Closed enterprise research preview; no self-serve endpoint, access via request form |
| Release Date | November 15, 2026 (research preview) |