{"id":"moshi","name":"Moshi","maker":"Kyutai","url":"https://github.com/kyutai-labs/moshi","repo":"kyutai-labs/moshi","package":"pypi:moshi","summary":"A speech-text foundation model and full-duplex spoken dialogue framework that listens and speaks at the same time.","details":"Provides model code and weights, with inference in Python, MLX and Rust. It models both sides of a conversation so it can handle interruptions. The code is Apache 2.0 licensed.","open_source":true,"hosted":false,"verified":"2026-11-15","category":"voice","category_name":"Voice agents and realtime speech","page":"https://modelmoot.com/tools/moshi","reviewed":"2026-11-15","checked":{"at":"2026-11-15T15:07:43.409Z","address_answers":true},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}