# How to Break Down and Rewrite a RedNote (Xiaohongshu) Post: From Structure to Your Own Image Post Package

> Save Xiaohongshu posts to Beav, click Decompose to get the topic and structure template (plus shots, a per-shot script and clips for video posts), then use AI rewrite to keep the reference's text-to-image relationships while swapping in your own content, ending with a ready-to-publish image post package. Steps, plus how it compares with general AI chat and Xiaohongshu analytics platforms.

- Source: https://beav.pro/en/blog/features/xiaohongshu-breakdown-rewrite
- Author: JambaHailar (https://x.com/JambaHailar), Beav
- Published: 2026-10-23
- Last updated: 2026-10-23

> **Key takeaways**
>
> * Save a RedNote post to Beav with the browser extension, open it and click **Decompose**. Image posts give you a topic and a structure template; video posts that are downloaded to your computer also give you shots, a per-shot shooting script and a timed transcript.
> * **AI rewrite** opens a new chat that already references the post. Say what to keep and what to change; the AI first explains the structure it saw, then, once you agree, builds a Xiaohongshu image post package: title, caption, tags and every image.
> * Decomposition uses model quota, the rewrite uses your chat model, and images use image generation. You are borrowing structure and style; the facts, experiences and numbers must be yours.

Most people with a folder full of viral RedNote posts do the same thing when it is their turn to write: drop a screenshot into an AI chat and say "write me one like this." The result is usually one of two things. Either it drifts so far that whatever made the original work is gone, or it swaps a few words and lands uncomfortably close to copying.

What is missing is the middle step: working out how the original is put together, then deciding which relationships to keep and which to replace. Beav splits the job into three parts, save, break down and rewrite, and keeps the result of each in the knowledge base and drafts on your own computer.

The writing craft itself, such as filling in a structure card and editing out the AI voice, is covered in [viral post breakdown and rewriting](https://beav.pro/en/blog/workflows/viral-post-breakdown-rewriting). This article is about doing it in Beav.

## Who it is for

* Xiaohongshu creators with a set of benchmark posts who want a steady output of image or video posts.
* Marketers running several accounts who need to explain to a team why a post works.
* Anyone who wants to move a proven post structure onto their own product, store or field.

## Step 1: Save the post

The full walkthrough is in [how to save a whole RedNote post to your computer](https://beav.pro/en/blog/features/save-xiaohongshu-videos-and-images). Here are only the points that matter for breakdown and rewriting.

![Three points for saving a RedNote post: open it on your computer, save with the extension and pick the options, wait for media to download](https://beav.pro/blog/features/xiaohongshu-breakdown-rewrite-steps-en.webp)

*A complete save gives the breakdown and the rewrite something to work with.*

1. **Open the post page on your computer.** Use Chrome or Edge with the Beav extension and let the images or video load.
2. **Save with the extension and pick two options.** **Save comments** captures the comments too. **Decompose after saving** queues a breakdown as soon as the save succeeds. Both are off by default.
3. **Wait for the media in the knowledge base.** "Downloading" on the card means media is still arriving in the background. A video post can only be cut into shots once the video is on your computer.

Found the post on your phone? [Phone and computer pairing](https://beav.pro/en/blog/tutorials/beav-phone-computer-transfer) hands the link to your computer. You can also paste the link under **+ → Add link** in the knowledge base; that route goes through the capture service, which needs sign-in or a configured capture API and consumes quota according to its rules.

## Step 2: Break it down

Open the post in the knowledge base and click **Decompose**. For many posts at once, select them in the list, right-click, and confirm the item count and model usage; a batch takes up to 100 items. Decomposition uses model quota.

![What image posts and video posts produce when decomposed: topic and structure template, plus shots, a per-shot script, a timed transcript and clips for video](https://beav.pro/blog/features/xiaohongshu-breakdown-rewrite-outputs-en.webp)

*Image posts give you topic and structure; video posts add shot-level results.*

**Image posts** mainly give you a topic and a structure template. The topic also goes to the topic center, so you stop digging through bookmarks for ideas. Read the template next to the original images: what the cover and title each do, where the text turns, and what the ending asks the reader to do. Image posts do not produce shots or clips, so set your expectations accordingly.

**Video posts** that are downloaded to your computer also give you shots, a per-shot shooting script and a timed transcript, and the keyframes, shot MP4 clips and audio cues are copied to `decompositions/<title>` in the asset library. If a cut is off, merge a shot into the previous one or split at the playback position, then decompose again. The details are in [how to break down a viral short video](https://beav.pro/en/blog/features/video-breakdown-shot-script).

If you saved the comments, **Comment insights** appears next to the comment count in the post details and lets the AI analyze readers' questions and disagreements; see [Xiaohongshu comment insights](https://beav.pro/en/blog/features/xiaohongshu-comment-insights). The breakdown tells you how the post was written, the comments tell you what readers still want, and together they give your next topic a reason to exist.

Do not jump straight to rewriting. Spend two minutes with the result and the original and ask: which one or two things made this post work? That call is yours to make, and it is what lets you say clearly what to keep.

## Step 3: Rewrite it into an image post package

Click **AI rewrite** at the top right of the post details. Beav opens a new chat with a prefilled rewrite prompt, a reference to this post and the rewriting skill selected. Nothing is sent until you send it, so add your requirements first.

Spell out three things:

1. **What to keep.** For example: "Keep the big-text cover with a side-by-side comparison image, the reading order of the first three pages, and one conclusion per page."
2. **What to change.** The topic, product, people, audience, and every specific fact.
3. **What you want back.** Ask to "make it a Xiaohongshu image post" and by default you get a complete image post package: title, caption, tags and every image. If you only want the copy or a plan first, say so.

The AI then explains how the original is organized, based on the text and images it actually read, and proposes how the new post maps onto it. By default it waits for you to approve the plan. This is what the docs call keeping the relationships that make the reference work: which label points at which object, whether two things are compared on the same page, which line goes with which image. When the product changes, the labels, dialogue and pictures change with it, not just the caption.

Without permission to generate every image at once, the AI makes two pages first for you to check, then continues. Check each page for text, facts and layout in your drafts. For small fixes, say "change only page two" or "swap these two pages"; the AI edits the same package and leaves unchanged images alone. Xiaohongshu post text is capped at 1,000 characters.

Two cautions. Numbers and before-and-after images in the original are not facts about your product, so supply your own evidence. And if images could not be read or pages are missing, the AI should tell you what it actually saw, which tells you how far to trust the plan. Once the package is right, save it to your computer or pick it up on your phone by scanning a code. You publish it yourself; Beav never posts for you.

## How it compares with the usual approaches

![Asking AI to write one like this compared with breaking the post down first and keeping its relationships](https://beav.pro/blog/features/xiaohongshu-breakdown-rewrite-contrast-en.webp)

|                              | Screenshot into a general AI chat   | Xiaohongshu analytics platforms (e.g. Qiangua, Xinhong)               | Beav                                                                      |
| ---------------------------- | ----------------------------------- | --------------------------------------------------------------------- | ------------------------------------------------------------------------- |
| Starting point               | A screenshot or pasted text         | The platform's library of hit posts, keywords and creator data        | Posts you saved yourself, with original images, video and comments        |
| What the breakdown gives you | A written analysis you can question | Engagement data, trends and rankings                                  | Topic and structure template; video posts add a per-shot script and clips |
| Rewrite result               | Text; images are a separate job     | Depends on whether the platform offers AI copywriting; check each one | An image post package: title, caption, tags and all images                |
| Strengths                    | Always at hand, flexible questions  | Finding hits, industry data and ad references                         | Breakdown to finished post in one place, material kept locally            |
| Cost                         | Each service's plan                 | Usually a paid membership                                             | Model quota for decomposition; chat and image model quota for the rewrite |

Analytics platforms are built for finding things: which posts are rising, which creators are running campaigns. Beav does not do that. A general AI chat wins on convenience and is great for talking through why a post works. Beav is for the moment you already know which posts to learn from and want that structure to become a post you can actually publish.

## Limits and notes

* Image posts produce a topic and a structure template only, with no shots or clips. Video posts need the video on your computer before shots can be cut.
* Shared spaces do not offer decomposition.
* **AI rewrite** only prepares the chat. Read the prompt before you send it, and check every page of the result.
* You are learning structure and style. The author's experiences, test results and unique conclusions are not yours to reuse. Get the author's permission before using their images or video clips.
* Beav does not publish for you and does not promise watermark removal.

## Next steps

* Build up benchmark posts first: [batch capture Xiaohongshu notes](https://beav.pro/en/blog/features/batch-capture-xiaohongshu-notes).
* Have new posts from benchmark creators broken down automatically: [batch and automatic decomposition](https://beav.pro/en/blog/features/batch-breakdown-and-auto-breakdown).
* Full reference: [from capture to draft](https://beav.pro/en/docs/workflows/collect-to-create#keep-the-relationships-that-make-the-reference-work) and [drafts, packages and publishing](https://beav.pro/en/docs/workspace/drafts-and-publishing).

*Applies to Beav desktop 2.9 and browser extension 2.7. Checked 2026-10-21.*