abus-aikorea / voice-pro

Gradio WebUI for creators and developers, featuring key TTS (Edge-TTS, kokoro) and zero-shot Voice Cloning (E2 & F5-TTS, CosyVoice), with Whisper audio processing, YouTube download, Demucs vocal isolation, and multilingual translation.

m-bain / whisperX

WhisperX: Automatic Speech Recognition with Word-level Timestamps (& Diarization)

olewave / fa-bench

Phone and word timestamp accuracy on human-labeled TIMIT and Buckeye, clean and under reverb, noise, music and babble. Two tracks, forced aligners given the transcript and ASR systems that find their own words. MFA, WhisperX, Charsiu, Qwen3, Parakeet-TDT, Olign and more. Monthly leaderboards, paper on arXiv.