Frame QA Studio: mark and fix storyboard frame errors

Frame QA Studio is a review interface for AI-generated storyboard frames. A reviewer steps through a sequence of frames, flags what is wrong in each one using five error categories (character, emotion, prop, composition and segments) with keyboard hotkeys, and adds a note. A second view collects every flagged frame into a fix queue, where the reviewer records what needs to change, why the error occurred (prompt issue, model limitation or data gap) and writes the corrected prompt text for regeneration.

The problem and who it is for

The problem it addresses: when an explainer-video pipeline generates hundreds of frames, QA has to be faster than the generation itself, and the corrections need to feed back into the prompts rather than get lost in chat threads. It is for the people reviewing frames, a producer or a writer, not for the model engineer.

Stack and status

Stack observed in the source: React 18 with in-browser JSX, no build step and no backend. Status: an interaction prototype, not a shipped tool. It is one of the explainer-video tools listed on the homepage.

Live demo

Prototype status: interaction demo · all 20 frames, four characters, six props and every detection score are mock data generated in the page

Use ① Mark Errors to flag frames, then ② Fix Errors to write corrections. The demo needs JavaScript and a desktop-width window.