Term
Chatterbox
Chatterbox is an open-source text-to-speech model from Resemble AI under the MIT license. In blind tests most listeners preferred it over ElevenLabs; it offers emotion control, voice cloning and low latency.
Chatterbox — explained in more detail
Chatterbox is an open-source text-to-speech model from the provider Resemble AI, released in 2025 under the permissive MIT license. This allows the weights and code to be used, adapted and self-hosted freely. In blind tests a majority of listeners preferred Chatterbox over the commercial ElevenLabs, praising naturalness, clarity and emotional range.
Its features include emotion control, voice cloning from a few seconds of reference audio, multilingual output and low latency in the range of a few hundred milliseconds. Because the model is openly available, it suits operation on your own infrastructure without vendor lock-in; it reached wide adoption on GitHub and Hugging Face.
Example / In practice
A team that wants to run speech output in its own data center in a privacy-compliant way can set up Chatterbox locally, clone a company voice from a short recording and use it to voice announcements or product videos without sending text to an external service.
Distinction from similar terms
Unlike the closed ElevenLabs v3, Chatterbox is freely licensed as an open-weight model and can be self-hosted. Like other speech and voice models it generates speech from text (TTS) and is distinct from speech recognition; its profile centers on open availability while remaining competitive in quality.
Discover more
AI in Review — August 2026: A Price U-Turn, the Open-Weight Wave and an Enforceable EU AI Act
August 2026 in AI: Sonnet 5 price freeze, Qwen3.8-Max, the open-weight wave, Google AI Mode on Gemini Flash and the enforceable EU AI Act — sorted out.
GlossaryFish Audio S2
Fish Audio S2 is an open-source text-to-speech model from the Fish Speech family. With inline tags it controls emotion and emphasis at word level, supports more than 80 languages and delivers very low latency.
EncyclopediaFLUX.2
FLUX.2 by Black Forest Labs (Freiburg): open-weight image model, variants from 4B to 32B, licenses, hardware needs and how it stacks up in the market.