How nzeronfourme/movieshort-ai works
MovieShort AI turns a movie into vertically formatted Shorts through a Gradio interface: analyze scenes and dialogue, select clips (with optional LLM scoring), then render crops and subtitles. Batch processing owns the analysis and clip-selection flow; the rendering pipeline handles per-clip output. The graph also includes settings, provider integrations, and person tracking. Relationships are limited to README statements and import/call references in the supplied SOURCE INDEX or excerpts; unsampled implementation details are not inferred.
Main entry points: pipeline.py Render pipeline · app.py Gradio interface · batch.py Batch selection · ffmpeg_utils.py Video transforms
Components 18
User interface
- Gradio interface
gui/app.py
Movie analysis
- Batch selection
core/batch.py - Scene detector
analyzers/detector.py - Scene analysis
analyzers/scene_analyzer.py - Clip scoring
analyzers/text_analyzer.py
Clip rendering
- Render pipeline
core/pipeline.py - Person tracking
core/processor.py - Transcription and subtitles
core/subtitle.py - Video transforms
utils/ffmpeg_utils.py - Font management
utils/font_manager.py
Configuration and utilities
- Runtime defaults
config.py - User settings
utils/user_config.py - Log capture
utils/log_capture.py
External services
- LLM providers
- FFmpeg tools
- faster-whisper
Other components
- Movie editor
- Rendered clips
output
Connections 26
- Movie editor → Gradio interface: selects movie
README.md - Gradio interface → Batch selection: starts batch
gui/app.py - Gradio interface → Render pipeline: starts manual render
gui/app.py - Gradio interface → User settings: loads and saves
gui/app.py - Gradio interface → Clip scoring: verifies provider key
gui/app.py - Gradio interface → Log capture: captures logs
gui/app.py - Batch selection → Scene detector: requests scenes
core/batch.py - Batch selection → Scene analysis: uses analysis
core/batch.py - Batch selection → Clip scoring: scores blocks
core/batch.py - Batch selection → Transcription and subtitles: transcribes movie
core/batch.py - Batch selection → Render pipeline: renders selected clips
core/batch.py - Scene detector → Scene analysis: uses scene analysis
analyzers/detector.py - Clip scoring → LLM providers: sends scoring requests
analyzers/text_analyzer.py - Scene analysis → FFmpeg tools: probes and extracts media
analyzers/scene_analyzer.py - Transcription and subtitles → faster-whisper: transcribes audio
- Render pipeline → Person tracking: tracks people
core/pipeline.py - Render pipeline → Transcription and subtitles: generates subtitles
core/pipeline.py - Render pipeline → Video transforms: applies video transforms
core/pipeline.py - Render pipeline → Font management: resolves subtitle font
core/pipeline.py - Render pipeline → User settings: loads preferences
core/pipeline.py - Video transforms → Font management: uses font assets
utils/ffmpeg_utils.py - Video transforms → FFmpeg tools: runs media commands
utils/ffmpeg_utils.py - Render pipeline → Rendered clips: writes rendered clips
README.md - Batch selection → Runtime defaults: uses defaults
core/batch.py - Render pipeline → Runtime defaults: uses defaults
core/pipeline.py - Gradio interface → Runtime defaults: reads runtime settings
gui/app.py