Video QA and continuity checklist
Catch the artefacts, continuity breaks, brand errors and rights issues in AI-generated video before your audience does.
Fill in your details
Values are inserted into every prompt below. Nothing you type leaves your browser.
2 field(s) left — unfilled variables stay as {{placeholders}}.
1Per-shot artefact check (watch each shot at 0.5x speed)
☐ Hands and fingers: count, shape, how they hold objects. ☐ Faces: no melting, flicker or identity change mid-shot. ☐ Physics: objects don't pass through each other, liquids and fabric behave plausibly. ☐ Background: no morphing buildings, people appearing or vanishing. ☐ Text and logos: none garbled (replace with editor text). ☐ Edges: no shimmering outlines or warping at the frame border. ☐ First and last frames: no sudden jumps.
2Continuity check (watch the full cut)
☐ Character wardrobe, hair and accessories match across shots. ☐ Props are the same colour, size and position across consecutive shots. ☐ Light direction and time of day are consistent within a scene. ☐ Screen direction: a character moving left-to-right keeps doing so across cuts unless intended. ☐ Colour grade matches across generated and filmed footage. ☐ Audio: ambient sound doesn't jump between shots in the same location.
3Brand, claims and rights
☐ Product depiction matches the real product (colour, shape, features). ☐ Every claim on screen or in voice-over is substantiated. ☐ No real person's likeness, recognisable artwork or third-party trademarks without permission. ☐ Music and voice licensed for commercial use. ☐ Platform AI-content disclosure labels applied where required. ☐ Captions present and accurate.
4AI-assisted review
Multimodal models that accept video uploads can do a first-pass review. Treat their findings as leads to check by eye, not as a pass.
Review the attached video against the intended shot list and brand rules. Intended shots: {{shot_log}} Brand rules: {{brand_rules}} For each shot give a timestamp range and list: visual artefacts (hands, faces, physics, warping, text), continuity issues compared with neighbouring shots, deviations from the intended shot, and brand-rule violations. Rate each issue Minor / Fix before publishing / Blocker. Say explicitly when you are uncertain.
Example input
30-second bottle ad, 7 shots, brand rules on product colour and logo placement.
Example output
Shot 4 (0:12-0:16) — Fix before publishing: cyclist's jacket changes from yellow to orange around 0:14. Shot 5 (0:16-0:20) — Blocker: the bottle appears with a silver lid, not bamboo. Shot 6 (0:20-0:24) — Minor: background car briefly merges with a lamppost (uncertain; check at 0.25x).
Troubleshooting
- The model misses obvious artefacts: review shot by shot with short clips instead of the whole video.
- File too large to upload: export a lower-resolution review copy.
Known limitations
- AI review is a first pass only; a human must watch the final cut in full.
- Rights and claims checks require human judgement and, for regulated products, legal review.