rehearsal-studio/README.md

Rehearsal Studio

Desktop-first script rehearsal application for synchronizing screenplay text with one or more audio files, then practicing a selected role against that synchronized timeline.

Current implementation

• src/: React + TypeScript UI with import, reader, sync editor, provider routing, and practice views
• src-tauri/: Rust host for project management, media preparation, JSON artifacts, and SQLite-backed practice history
• providers/python/: provider runtime for script parsing, draft transcription, fuzzy alignment, and transcript-based answer evaluation

The provider layer is modular by design:

• local-whisper-draft is the default local provider
• remote-placeholder is reserved for future API adapters

Local development

1. Install JavaScript dependencies: npm install
2. Install Python provider dependencies if you want local document parsing and optional ASR: python -m venv .venv && .venv/bin/pip install -r providers/python/requirements.txt
3. Run the browser preview: npm run dev
4. Run the desktop shell: npm run tauri dev

Notes

• If faster-whisper is not installed or no local model is available, the provider runtime falls back to generating a draft transcript from script lines and total audio duration.
• The app automatically prefers ./.venv/bin/python or ./.venv/Scripts/python.exe for the provider runtime. REHEARSAL_PYTHON still overrides that if you want a different interpreter.
• Media preparation requires ffmpeg and ffprobe on the system path.
• The UI accepts absolute file paths for the first implementation; native file-picker integration can be added on top of the existing commands.