# How to Break Down a Viral Short Video: Topic, Structure, Shot-by-Shot Script and Clips in One Click

> Open a saved video in Beav and click Decompose to get the topic, a structure template, shots, a per-shot shooting script and a timed transcript, with keyframes, shot MP4s and audio cues copied to your asset library. Steps, model usage, and how it compares with manual shot analysis and general AI chat.

- Source: https://beav.pro/en/blog/features/video-breakdown-shot-script
- Author: JambaHailar (https://x.com/JambaHailar), Beav
- Published: 2026-11-01
- Last updated: 2026-11-01

> **Key takeaways**
>
> * Open a saved video in the Beav knowledge base and click **Decompose**. You get a topic, a structure template, a list of shots, a per-shot shooting script and a transcript with timestamps.
> * It is more than text: the transcript, keyframes, shot MP4 clips and audio cues are copied to the asset library, and the topic goes to the topic center, ready for your own videos.
> * If a cut is off, merge a shot into the previous one or split at the playback position, then decompose again to get clips at the new cuts. Decomposition uses model quota.

Studying a video shot by shot is the most reliable way to learn short-form video: watch frame by frame and note what each shot shows, what is said and how long it lasts. It is also slow. A one-minute Douyin or RedNote video can take half an hour, plus the screenshots, the clip trimming and the transcript.

Beav's breakdown does that groundwork for you: **it cuts a video into shots, gives each shot a description, its lines and its timing, and saves the clips and keyframes as reusable material**. Your job is to understand why it was shot that way.

Breaking down and rewriting image posts is covered in [viral post breakdown and rewriting](https://beav.pro/en/blog/workflows/viral-post-breakdown-rewriting). This article is about video only.

## Who it is for

* Creators making talking-head, review, unboxing or tutorial videos who want to learn a benchmark video's pacing.
* Video leads and marketers who hand shot lists and scripts to a team.
* Anyone who wants to keep good shots and music as editing references.

## How to use it

First save the video to your local knowledge base, for example by clicking "+" in the knowledge base, choosing **Add link** and pasting a RedNote or Douyin video link. The video must be downloaded to your computer before shots and clips can be extracted.

![Four steps to break down a video: save it, click Decompose, check the shot script and cuts, take clips from the asset library](https://beav.pro/blog/features/video-breakdown-shot-script-steps-en.webp)

*Save the video, decompose it, adjust the cuts, take what you need from the asset library.*

1. **Open the video and click Decompose.** If the card still says "Downloading", wait for the media to finish.
2. **Read the five parts.** Topic, structure template, shots, per-shot shooting script and timed transcript. The shooting script says what each shot shows and how it is filmed; every transcript line carries its time, so you can match words to pictures.
3. **Check the shots against the video.** Click a shot thumbnail or a timestamp to jump there. If two shots are really one, merge the shot into the previous one; if one shot hides a clear cut, play to that point and split at the playback position. Later decompositions keep your edited cuts.
4. **Take the clips from the asset library.** Copies go to `decompositions/<title>` in the asset library: the transcript, keyframes, shot MP4 clips and M4A audio cues. The origin panel keeps the source video, author and time range, and the topic is added to the topic center.

After you adjust the cuts, decompose again to generate new clips; the old ones stay. Edits you made in the asset library are kept, and changed outputs are saved as new copies.

![What one video breakdown produces: topic, structure template, shot script, timed transcript, keyframes, shot MP4s, audio cues](https://beav.pro/blog/features/video-breakdown-shot-script-outputs-en.webp)

*Text results stay on the entry; clips and keyframes are copied to the asset library.*

## How it compares with manual analysis and general AI chat

|                     | Manual shot-by-shot analysis               | A general AI such as Gemini or ChatGPT          | Beav Decompose                                           |
| ------------------- | ------------------------------------------ | ----------------------------------------------- | -------------------------------------------------------- |
| What you get        | Your own notes and screenshots             | A written analysis you can question             | Topic, structure template, shot script, timed transcript |
| Clips and keyframes | Trim and capture them yourself             | Usually text only                               | Shot MP4s and keyframes saved to the asset library       |
| Adjusting cuts      | Entirely up to you                         | Describe it again and re-ask                    | Merge or split, and the next run keeps your cuts         |
| Strengths           | Deepest understanding; trains your own eye | Flexible follow-up questions about why it works | Results become material for topics and editing           |
| Cost                | Your time                                  | Each service's own plan                         | Model quota                                              |

Doing it by hand builds judgment that no tool replaces, and a general AI is great for open-ended questions about a single video. Beav's breakdown works more like an assistant editor: it handles the cutting, transcribing and frame grabbing, and leaves the results in your asset library.

## Limits and notes

* Transcription, topic and structure analysis, and the shooting script use transcription, chat and video-understanding models, and all of them use model quota. If one step fails, finished parts are kept; check the reason at the bottom of the result.
* Video over the video-understanding size limit is compressed to a low-resolution copy first. If it is still too large, the shooting script is skipped and the reason is shown.
* Audio cues come from stretches with sound but no speech, so they may include ambient noise. This is not vocal separation and does not split voice from music.
* Shared workspaces do not offer Decompose. Clips and music are for study; confirm reuse rights before publishing or commercial use.

## Next steps

* To decompose many videos at once, read [batch breakdown](https://beav.pro/en/blog/features/batch-breakdown-and-auto-breakdown).
* Save the video first: [save Douyin videos on desktop](https://beav.pro/en/blog/features/save-douyin-videos-on-desktop).
* Full reference: [decompose videos and notes](https://beav.pro/en/docs/workflows/collect-to-create#decompose-videos-and-notes). For managing the results, see [asset management](https://beav.pro/en/asset-management).

*Applies to Beav desktop 2.9. Checked 2026-10-21.*