Video QA and rough-cut factory
The team has many user case videos and needs to filter by aspect ratio, duration, speech quality, and content quality, then keep qualified videos and optionally generate rough-cut collections.
Pain points
Manual video review is slow and criteria are hard to keep consistent. / Qualified videos, review-needed videos, and rejection reasons must be delivered clearly.
Solution
Users upload a video folder, apply editable screening criteria step by step, use ASR transcripts for expression quality checks, and review outputs with preview, status edits, deletion, and package download.
Architecture
The flow combines local video metadata checks, audio extraction, ASR, screening rules, optional rough-cut tooling, and Excel/folder artifacts.
Business value
Automates first-pass video screening and delivery sorting so reviewers focus on cases that need human judgment.