How kirat11x/video--clipper-vizard-ai-replica-automatic-short-form-clip-generator- works
video-clipper is a local video-to-shorts pipeline with a Streamlit studio and CLI. Users provide a URL or file; the system ingests media, extracts audio and frames, analyzes speech, audio, semantics, hooks, and visuals, fuses signals into windows, ranks candidates, and renders platform-ready clips. The studio adds interactive scoring, clip editing, and batch rendering. The graph follows documented workflow and shown calls; some CLI orchestration details and UI-to-worker wiring are unsampled, so uncertain links are omitted.
Main entry points: run_pipeline.py Signal Fusion · rank_clips.py Clip Ranking · video_ingestion.py Media Ingestion · clip_engine.py Studio Engine
Components 19
Studio and Jobs
- Clipper Studio
app.py - Studio Engine
clip_engine.py - Background Jobs
job_runner.py
Media Analysis
- Media Ingestion
video_ingestion.py - Speech Transcription
audio_transcription.py - Audio Signals
audio_analysis.py - Text Semantics
text_analysis.py - Hook Scoring
hook_detector.py - Visual Signals
visual_analysis.py - Signal Fusion
run_pipeline.py
Clip Selection
- Clip Ranking
rank_clips.py - Legacy Selection
clip_selector.py - Audio Validation
audio_validator.py - Pipeline Evaluation
evaluate_pipeline.py - Genre Weights
weights.json - Suggested Clips
suggested_clips.json
Rendering
- Shorts Renderer
video_renderer.py - Rendered Shorts
Other components
- Video Creator
Connections 22
- Video Creator → Clipper Studio: provides source
README.md - Video Creator → Media Ingestion: provides source
README.md - Clipper Studio → Studio Engine: uses
app.py - Studio Engine → Clip Ranking: uses profiles
clip_engine.py - Media Ingestion → Speech Transcription: extracts audio for
README.md - Media Ingestion → Audio Signals: extracts audio for
README.md - Media Ingestion → Visual Signals: extracts frames for
README.md - Speech Transcription → Text Semantics: supplies transcript
README.md - Text Semantics → Signal Fusion: supplies semantics
README.md - Audio Signals → Signal Fusion: supplies signals
README.md - Hook Scoring → Signal Fusion: supplies hook scores
README.md - Visual Signals → Signal Fusion: supplies signals
README.md - Signal Fusion → Clip Ranking: writes signal windows
README.md - Genre Weights → Clip Ranking: configures scoring
README.md - Clip Ranking → Suggested Clips: writes candidates
README.md - Studio Engine → Shorts Renderer: uses for rendering
clip_engine.py - Background Jobs → Studio Engine: uses job logic
job_runner.py - Background Jobs → Shorts Renderer: runs render job
job_runner.py - Suggested Clips → Shorts Renderer: provides clip ranges
README.md - Shorts Renderer → Rendered Shorts: renders
README.md - Legacy Selection → Audio Validation: validates candidates
clip_selector.py - Pipeline Evaluation → Clip Ranking: imports ranking logic
evaluate_pipeline.py