Voice AI companion startup targeting consumers and developers. Its CSM (Conversational Speech Model) is an end-to-end transformer that generates expressive speech from conversational context rather than piping text through a separate TTS system. The 1B-parameter variant is open-sourced on Hugging Face under Apache 2.0. Smart glasses planned for 2027.
Sesame's open weights Conversational Speech Model pairs a Llama backbone with a Mimi audio decoder to produce dialogue that tracks context turn by turn.