This is a reference page on the BestTTS directory for KittenML's Kitten TTS 2, a 1.7B-parameter speech model, with an embedded demo of KittenML's official Hugging Face Space. The page summarizes the publisher's documentation: voice cloning from a 5 to 30 second sample, 123 named voices across 16 languages, CPU inference, and setup with the kittenml pip package. It also covers the Stellon Labs Community License, which requires registration for commercial use, and notes that the text normalizer is tuned for English. BestTTS does not provide the model, and the demo is a third-party embed hosted by KittenML, limited to 500 characters and 4 requests per minute.
New TTS voices and models launch constantly, and the facts about them are scattered across model cards, repos and pricing pages. BestTTS collects those facts in one place, including license terms.