✕
Download any video from YouTube, TikTok, Douyin, or any URL. Auto-transcribe it with WhisperX. Shadow sentence by sentence. Get instant AI pronunciation scores — all on your device, no cloud required.
YouTube, TikTok, Douyin, Twitter/X, Instagram — or any local file. No curated library, no lock-in. If yt-dlp can download it, ShadowStudio Desktop can transcribe and score it.
Transcription runs entirely on-device with WhisperX. Your recordings never leave your machine. Only the pronunciation scoring step (optional) contacts Azure Speech — nothing else.
Paste any video link and go. The entire pipeline runs automatically.
Paste any URL. Pick quality. ShadowStudio downloads the video and subtitles automatically.
WhisperX splits the video into sentences with millisecond-accurate timestamps — on your device.
Listen to one sentence, record yourself saying it. Adjust speed, replay, repeat.
Get instant phoneme-level pronunciation feedback from Azure Speech or a local whisper model.
Every screen from first launch to scoring your pronunciation.
Detects missing dependencies and installs everything automatically — Homebrew, ffmpeg, WhisperX, and your chosen transcription model.
Built for serious language learners who practice every day.
YouTube, TikTok, Douyin, Twitter/X, Instagram, and any yt-dlp supported platform. Subtitles auto-downloaded when available.
WhisperX with forced alignment for millisecond-accurate sentence timestamps. Supports multilingual and English-only models. No internet needed.
Sentence-by-sentence playback with adjustable speed (0.5× – 1.0×), keyboard shortcuts for hands-free control, and side-by-side comparison playback.
Azure Speech delivers phoneme-level accuracy, fluency, and completeness scores with color-coded word highlighting. Free mode uses whisper-cli offline.
Hands-free auto-advance through every sentence in the video. Perfect for intensive practice sessions without touching the keyboard.
English, Japanese, Chinese, Korean, French, German, Spanish, Portuguese, Italian, Russian, Arabic, Hindi, and more — powered by multilingual Whisper models.
Fine-tune timestamps with ±50ms adjustments, merge fragments, split long segments, and edit transcription text directly in the studio.
Preview and watch any library video without leaving the app. Full playback, seek, and keyboard controls — jump straight from download to watching.
All data stored in a local SQLite database. Transcription is fully on-device. Your videos and recordings never leave your machine.
ShadowStudio Desktop is open-source and free. No account required. The setup wizard handles everything on first launch.
macOS · Windows · Free
Apple Silicon & Intel · macOS 11.0+ · ~150 MB
Windows 10/11 x64 · ~180 MB