Open-source credits

AudioHash is built on open models and open infrastructure. Thank you to their authors.

• MeloTTS by MyShell.ai and MIT, MIT License, served on Cloudflare Workers AI.
• Kokoro-82M by hexgrad, Apache 2.0 (coming soon). Training data includes CC BY audio from Koniwa (CC BY 3.0) and SIWIS (CC BY 4.0).
• Aura-2 premium voices by Deepgram, served on Cloudflare Workers AI.
• lamejs MP3 encoder (a JavaScript port of LAME), LGPL-3.0, used unmodified on our servers to encode audio. Source: github.com/zhuker/lamejs.