wildminder/awesome-ai-voice

What it solves

Finding a good open-source text-to-speech (TTS) or voice generation model is hard because there are many options with different capabilities, languages, and licenses. This project gathers them in one place so you can quickly scan and compare the latest releases without hunting across GitHub and Hugging Face.

How it works

The repository is a single curated Markdown file. It starts with a table of contents and then a quick-comparison table for TTS models, listing features like voice cloning, ASR support, language coverage, streaming, and license type. Below that, each model gets its own expandable section with a description, release date, key characteristics, feature flags, and links to official resources (Hugging Face, GitHub, demos, and sometimes additional tools like ComfyUI nodes). The list is sorted by release date, newest first, and also covers categories beyond pure TTS: anything-to-audio, audio restoration and enhancement, speech recognition, audio codecs and tokenizers, and audio transcription models.

Who it’s for

Developers, researchers, and hobbyists who need to pick a TTS or voice generation model for their own projects, or who want to monitor what’s new in the open-source speech space. It’s also useful for comparing model capabilities and licensing before committing to one.

Highlights

• Sorted by release date so you always see the newest models first.
• Quick-comparison table lets you filter by voice cloning, ASR, languages, streaming, and license at a glance.
• Each model has a detailed profile with links to Hugging Face, GitHub, demos, and sometimes extra tools.
• Covers not only TTS but also audio restoration, speech recognition, codecs, and more.
• Includes both major models and niche community fine-tunes (e.g., an anime-style Japanese TTS).