# RedNote (Xiaohongshu) AI Asset Management: Turning a Messy Bookmark Folder Into a Local Knowledge Base > How creators move benchmark posts out of bookmarks, cloud drives and chat history into one local knowledge base, how to tag them, and why that library decides the quality of what AI writes for you. - Source: https://www.getbeav.com/en/blog/guides/local-ai-asset-management-knowledge-base - Author: JambaHailar (https://x.com/JambaHailar), Beav - Published: 2026-10-05 - Last updated: 2026-10-06 > **Key takeaways** > > * Saved material is only worth something if you reuse it. Saving a lot and never opening it is the same as not saving. > * The smallest unit to save is a complete post: cover, caption, image order, comments and source link, not a screenshot. > * Tags only need to answer four questions: which account it is for, what part is worth borrowing, what topic it belongs to, and whether it is still useful. > * Your library is the raw material for AI topics, drafts and images. The more specific the material, the less generic the AI output. Many creators keep material in several places: platform bookmarks, a cloud drive folder, screenshots in chat history. When you need "that post about home storage I saved last month", you often cannot find it. This post explains how to move all of that into one local knowledge base and keep it organized so you actually use it. ## Saving more does not help if you cannot find it The problem with bookmarks is usually not saving too little. It is that you never open them again. Bookmarks have no tags and poor search, so after a few weeks they disappear down the list. The test for a useful library is simple: do you look things up in it every week? Thirty posts you return to often are worth more than ten thousand bookmarks you never open. ## Save the whole post, not a screenshot A cover screenshot only tells you that you liked it at the time. Two weeks later you will not remember why it worked, and you cannot hand it to AI to analyze or rewrite. What is worth saving is the whole post: cover, caption, image order, the top questions in the comments, plus the source link. With the source link you can go back later and see whether new questions showed up in the comments. With the Beav browser extension, click "Save note" on a Xiaohongshu (RedNote) post. The caption, images, video, the comments visible on the page, and the like and save counts all go into your knowledge base together. For details, see [building a benchmark post pool with a browser extension](https://www.getbeav.com/en/blog/workflows/browser-extension-capture-pool). ## Few tags, useful tags Too few tags and you cannot find anything; too many and you cannot remember them. Have each item answer four questions: * Which account or line of business it is for * What is worth borrowing: the title, the opening, the image order, or the questions in the comments * Which topic or series it belongs to * Whether it is still useful; move outdated items out Why sort by account and topic? Because that is how you look for ideas: "what else can this account talk about?" and "how far have others taken this topic?" Leave the rest to search. The Beav knowledge base lets you add tags by hand and filter or search by tag, folder and platform. You do not need to re-tag all your old bookmarks; that takes too long. Fill in the few topics you use most, and save new material the right way from now on. ## Your library is AI's raw material This is the biggest change from a few years ago: what you give AI decides how good its output is. If you only say "write me a Xiaohongshu post about home storage", you get the generic content everyone has already written. If you ask it to work from the twenty storage posts you have broken down in your library, the angle, wording and structure become much more specific. The same goes for topics, drafts and covers. In Beav, material in the knowledge base can be used directly: each item has an "AI rewrite" button, and in a chat you can drag material into the input box as a reference. The complete workflow is in the [2026 Xiaohongshu AI marketing guide](https://www.getbeav.com/en/blog/guides/ai-xiaohongshu-operation-playbook). Good material still does not mean you can post the first draft. Fact-checking, opinions and judgment are still yours. ## Why keep it local Two reasons: privacy, and independence from platforms. Your list of benchmark accounts, unpublished topics and your own writing style are core assets for a content team. On your own computer, you know where they are and how they are used. Beav stores workspace data on your machine by default, and each account space has its own knowledge base. Platform bookmarks depend on the platform. If the rules change or the author deletes the post, the bookmark stops working. A complete post saved locally does not disappear when that happens. ## One thing to do today Pick the ten best benchmark posts in your niche, save each one completely into your local knowledge base, tag them with the four questions above, then ask AI to suggest five topics using only those ten posts. Compare that with the topics it usually gives you; the difference is obvious. The Beav desktop app runs on macOS, Windows and Linux and is on the [download page](https://www.getbeav.com/en/download). For an overview of the asset features, see [creator asset management](https://www.getbeav.com/en/asset-management).