Updated
Case evidence: Read the source breakdown report · View case frames and shot timings
Based on a published tool report. Production classifications are analytical judgments; suggested steps are not the original project or a verified recreation.
Start from shots, not from one prompt
The example video has 21 shots in 57 seconds: 15 live-action and 6 AI-generated. Compressing the whole ad into one prompt would mix two production routes that need different assets. The breakdown separates them first, then writes prompts only for the AI shots.
What changes between Seedance 2.0 and 2.5
The report’s plan states it plainly: the default engine, Seedance 2.0 (including Mini and Fast variants), cannot produce several shots in one call, so the 6 AI shots would be generated one by one. Seedance 2.5 can take multi-shot prompts of up to 30 seconds per call, which brings the whole AI portion down to 6 calls. Plan the call count against the engine version you actually have, and re-check current limits before committing.
Grouping rules from the real plan
Merge AI shots that sit in the same set and are joined by hard cuts into one generation. Keep each call at 30 seconds or less; keep hard cuts inside a unit under about 8 seconds (10 at most). Split at dissolves, fades and set changes — those transitions signal a new generation unit. These rules came from matching the edit structure to engine limits, not from guesswork.
What stays out of the prompt
The example ad carries roughly 300 caption items, dozens of text labels and stickers, all composited after generation. Captions, logos and stickers belong in editing where wording and timing stay under your control. Voice-over is better produced as separate TTS and mixed in, so you can change the script without regenerating picture. Use a first-frame image when a shot needs exact composition: generate that image first, then attach it to the call.
What this guide does not claim
The call plan and rules above come from a real published report, but rea-video does not generate videos, and we have not verified every prompt against a live Seedance run. Treat the prompts you derive as structured drafts, test one unit first, and adjust duration and motion to what your engine version returns.
A reusable analysis template
Call # / shots included / total duration Engine version and per-call limit First-frame reference (if used) Per shot: subject, framing, action start→end, camera, light Keep out: captions, logos, stickers, voice-over Split points: dissolves, fades, set changes
Practice with a real report
Download the ad’s report and find the remake plan: six generation units, the 2.0-versus-2.5 call comparison, and the first-frame rules for its closing shot. It is the source for every rule quoted above.
Open the case and download its report →