I’d like you to build a Windows-desktop application that uses AI to automatically align written lyrics with the vocal track in MP4 files and then lets me fine-tune every timestamp by hand. Core workflow • I load an MP4 (audio or video) and paste or import the raw lyrics text. • The app runs forced-alignment / speech-to-text under the hood to create initial time-codes. • A timeline editor appears where I can nudge, split, or merge lines and hear the changes instantly. Essential capabilities – Primary file type: MP4 (being able to drop in MP3, WAV or FLAC later is a bonus, but MP4 support is non-negotiable). – Full UTF-8, multi-language lyric handling so accents and non-Latin scripts display and export correctly. – Export options: LRC, SRT and plain-text with time-stamps; ideally also embed lyrics back into the MP4 metadata. – Clean, lightweight installer for Windows 10/11 that runs offline once the model is downloaded. Deliverables 1. Compiled .exe with installer 2. Source code and build instructions 3. Pre-trained model files or a script that fetches them automatically 4. Short README showing how to import, sync, edit, and export a song Acceptance criteria • Auto-sync places 90 % of lines within ±200 ms on a clear English test track. • Manual editor lets me adjust individual time-codes at single-frame resolution without noticeable playback lag. • Japanese lyric sample round-trips (import → sync → export) with characters intact. Open to Python (PySide/PyQt), C# (.NET), or C++/Qt—choose what you’re fastest with, provided the UI stays snappy. Let me know the speech/alignment libraries you plan to integrate so I can confirm license compatibility before you start.