Discover and explore top open-source AI tools and projects—updated daily.
rzruML-powered karaoke app with real-time scoring and dynamic visuals
Top 31.4% on SourcePulse
Nightingale is an ML-powered Karaoke app transforming any song into a karaoke experience. It targets music enthusiasts by automating vocal separation, lyric transcription with timestamps, and playback customization, simplifying karaoke creation from personal libraries or streaming sources within a self-contained binary.
How It Works
Nightingale employs a multi-stage ML pipeline using UVR Karaoke/Demucs for vocal isolation and WhisperX for word-level lyric transcription/alignment. It supports multiple ASR engines (Whisper, Parakeet v3) and alignment backends, including experimental GPU acceleration and CJK language support. Built with Tauri (Rust+React), it communicates with a persistent local analyzer via TCP sockets, ensuring ML model initializations occur only once for reduced per-song overhead and efficient caching.
Quick Start & Requirements
docs/self-hosted, docs/usdx available.Highlighted Details
Maintenance & Community
Developed by a single individual. Support via Patreon and Ko-fi. No specific community channels (Discord/Slack) mentioned.
Licensing & Compatibility
Licensed under GPL-3.0-or-later. This strong copyleft license may restrict linking with proprietary or closed-source software.
Limitations & Caveats
Linux lacks auto-updates; manual installation is needed. macOS requires manual Gatekeeper bypass (xattr -cr). UltraStar Deluxe support and some advanced alignment backends are experimental.
5 days ago
Inactive