Second-generation Octave TTS, released as a preview on October 1, 2025, with 11 languages, about 100 ms model latency and word and phoneme timestamps, though acting instructions were still pending. Hume ends TTS API access November 13, 2026.
Hume's third-generation Empathic Voice Interface, one speech-language model that transcribes, responds and speaks with prompt-designed or cloned voices. It remains the default EVI version alongside EVI 4-mini, but Hume closes the EVI API on November 13, 2026.
Hume's first Octave text-to-speech model, launched February 2025 as an LLM-based system that follows natural-language acting instructions and designs voices from text prompts, in English and Spanish. Superseded by Octave 2; available until Hume's TTS API closes November 13, 2026.
Octave request without a pinned version, for which Hume routes each call to whichever Octave model it judges most appropriate. Hume advises pinning version 1 or 2 for repeatable output; its TTS API shuts down November 13, 2026.
