Back to projects
Video automation
AI multimedia pipeline
A Windows application with GUI and CLI for analyzing video, transcribing, detecting scenes, and preparing outputs for multiple platforms.
- Responsibility
- Pipeline design, Python development, and Windows packaging
- Period
- 2026
- Client
- Viajando con el Pari
The problem
Reviewing hours of footage, finding strong moments, creating subtitles, and preparing rough cuts consumed too much repetitive editorial work.
The solution
I built a pipeline using FFmpeg, WhisperX, diarization, SceneDetect, candidate ranking, subtitles, platform variants, Premiere and Resolve exports, and publication tracking.
Outcome
Long-form material becomes review packages with candidates, frames, subtitles, editing guides, and interchange files ready for the editor.
What I owned
- Audio and video ingestion and normalization
- Transcription, diarization, and scene analysis
- Platform-specific candidate selection
- Subtitle and editorial exports
- GUI, CLI, analytics, and Windows executable
Deliverables
- Desktop application
- Command-line pipeline
- Subtitles and rough cuts
- Contact sheets and candidates
- CSV tracking and documentation
Technology
PythonFFmpegWhisperXpyannoteSceneDetectPyInstaller
