Prepare Amazon videos with AI
Inspect a video for its selected Amazon placement, review evidence frames and reports, and deliberately export a corrected MP4 when useful.
Before an upload, prepare Amazon videos with AI around the placement you actually intend to use. The amazon-video-prep skill inspects the file, extracts review frames and separates confirmed technical failures from warnings and unverified items.
If a product demonstration has the wrong aspect ratio, inspect the product and text before choosing a crop. Keep the source, make a separate export and compare its new report. The workflow prepares material for review; it neither uploads the video nor predicts Amazon’s moderation decision.
What it does
- Inspect dimensions, duration and encoding settings
- Keep technical failures separate from uncertain rules
- Review product visibility and mobile text
- Choose excerpts after inspecting their context
- Export a corrected copy without replacing the source
- Record remaining creative and audio questions
Install and use
- Provide the complete folder to a file-and-shell/vision-capable agent. Have Python 3, FFmpeg and FFprobe ready; Pillow is optional for contact sheets. Use authorised footage and product details.
- Specify the placement, destination store and content language, with a fresh output directory. Check current destination requirements instead of applying one generic Amazon-video profile.
- Review reports and sampled frames, then record creative findings with timestamps. Use more frames or available audio/transcripts where needed; unseen sections stay unverified.
- Approve any crop, excerpt or upscale explicitly. Inspect the corrected video and report, including product/text visibility and required provenance labels; rescaling cannot restore detail.
FAQ
Source
Made by Radu Bobirnac / Profit First PPC. Original skill: amazon-video-prep, published on GitHub under the MIT licence.
View on GitHub