Blog
Technical notes from the Tensorial team.
2026-09-30
Tensorial TTS v2: 43 languages, 29 more on trial, and a voice that starts in 70 milliseconds
Our multilingual speech model in production: what it speaks, how it streams, and how we separate what it learned from what it borrows.
Read more →2026-07-27
Tensorial T2 Live: a real-time multimodal conversation model
Our second generation: memory that survives the session, proactivity that knows when to speak, and eight languages.
Read more →