How jakob-ui/videogen_pipeline works
This repository is a Python prototype for producing a product video: the main script records a locally served webpage, generates a cloned German voice-over, then combines audio and video into an MP4. The documented web application, HTTP trigger API, and dynamic product-value updates are not implemented in the shown files; the current script uses fixed page and narration content. The graph reflects the sampled runtime code, not those planned capabilities.
Main entry points: pipeline.py Pipeline orchestration · voiceOver.py Voice cloning · Airspace_text_Ai.wav Clone reference audio
Components 10
Webpage recording
- Pipeline orchestration
Pipeline/pipeline.py - Webpage capture
Pipeline/pipeline.py - Recorded video
Voice-over generation
- Voice cloning
Pipeline/Audio/voiceOver.py - Clone reference audio
Pipeline/Audio/trainVoice/footage/Airspace_text_Ai.wav - Qwen TTS model
- Generated narration
Video assembly
- Audio-video muxing
Pipeline/pipeline.py - Product video MP4
Other components
- Script operator
Connections 11
- Script operator → Pipeline orchestration: runs
Pipeline/pipeline.py - Pipeline orchestration → Webpage capture: records page
Pipeline/pipeline.py - Webpage capture → Recorded video: produces
Pipeline/pipeline.py - Pipeline orchestration → Voice cloning: generates narration
Pipeline/pipeline.py - Voice cloning → Clone reference audio: clones from
Pipeline/Audio/voiceOver.py - Voice cloning → Qwen TTS model: loads and uses
Pipeline/Audio/voiceOver.py - Voice cloning → Generated narration: writes
Pipeline/Audio/voiceOver.py - Pipeline orchestration → Audio-video muxing: combines media
Pipeline/pipeline.py - Recorded video → Audio-video muxing: video input
Pipeline/pipeline.py - Generated narration → Audio-video muxing: audio input
Pipeline/pipeline.py - Audio-video muxing → Product video MP4: writes MP4
Pipeline/pipeline.py