rea-video.Video Breakdown Assistant
中文Get started
Production guides

YouTube video to prompt: analyze a reference shot by shot

A YouTube link is a starting point for research. To build a useful prompt, you need the actual frames, their timing and the layers added in editing.

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.

1. Start with a file you can analyze

This website does not offer a YouTube link-to-prompt upload form. Use your own export or a file supplied with permission, then follow the installation guide to run rea-video in your agent. Record the source URL and the selected time range alongside the file. A transcript alone cannot establish camera movement, framing or visual overlays.

2. Ask for a production breakdown

Request shot boundaries, visible action, camera movement, captions, audio cues and uncertain judgments. For a long reference, begin with the sequence you want to learn from. Keep both clip-relative and source-relative timecodes if you trim the file, so a collaborator can find the same moment.

3. Use the dialogue case as an output example

The linked dialogue report is a local-file example, not a tested YouTube import. Shot 2 runs from 00:01.375 to 00:02.833; shot 3 begins at 00:02.833. The report treats character footage and text overlays separately. Draft the character action for each shot, then place titles in the edit. Asking for the whole conversation, labels and cuts in one prompt loses that structure.

4. Keep claims and costs explicit

An inferred prompt is a new description, not the original creator’s prompt. Local preprocessing also does not guarantee offline analysis: a configured cloud model may receive frames, audio or video. Check your provider and configuration before processing private work; model usage can cost extra.

A reusable analysis template

Source URL / permission:
Local file / selected range:
Clip time → source time:
For each shot: action, framing, camera, duration
Separate: captions, dialogue, music, transitions
Output: draft prompt + required assets + uncertain claims

Practice with a real report

Compare shots 2 and 3 in the source report. Preserve their separate time ranges when drafting two prompts.

Open the case and download its report →

Continue your workflow