[{"data":1,"prerenderedAt":27},["ShallowReactive",2],{"i-lucide:grip":3,"i-lucide:chevron-right":8,"i-lucide:moon":10,"i-lucide:sun":12,"i-lucide:languages":14,"i-lucide:chevron-down":16,"i-lucide:shield-check":18,"i-lucide:mail":20,"blog-body-ai-xiaohongshu-cards-from-notes-en":22,"i-lucide:lightbulb":23,"i-lucide:pen-tool":25},{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":7},0,24,false,"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Ccircle cx=\"12\" cy=\"5\" r=\"1\"\u002F>\u003Ccircle cx=\"19\" cy=\"5\" r=\"1\"\u002F>\u003Ccircle cx=\"5\" cy=\"5\" r=\"1\"\u002F>\u003Ccircle cx=\"12\" cy=\"12\" r=\"1\"\u002F>\u003Ccircle cx=\"19\" cy=\"12\" r=\"1\"\u002F>\u003Ccircle cx=\"5\" cy=\"12\" r=\"1\"\u002F>\u003Ccircle cx=\"12\" cy=\"19\" r=\"1\"\u002F>\u003Ccircle cx=\"19\" cy=\"19\" r=\"1\"\u002F>\u003Ccircle cx=\"5\" cy=\"19\" r=\"1\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":9},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"m9 18l6-6l-6-6\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":11},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M20.985 12.486a9 9 0 1 1-9.473-9.472c.405-.022.617.46.402.803a6 6 0 0 0 8.268 8.268c.344-.215.825-.004.803.401\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":13},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Ccircle cx=\"12\" cy=\"12\" r=\"4\"\u002F>\u003Cpath d=\"M12 2v2m0 16v2M4.93 4.93l1.41 1.41m11.32 11.32l1.41 1.41M2 12h2m16 0h2M6.34 17.66l-1.41 1.41M19.07 4.93l-1.41 1.41\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":15},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"m5 8l6 6m-7 0l6-6l2-3M2 5h12M7 2h1m14 20l-5-10l-5 10m2-4h6\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":17},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"m6 9l6 6l6-6\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":19},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M20 13c0 5-3.5 7.5-7.66 8.95a1 1 0 0 1-.67-.01C7.5 20.5 4 18 4 13V6a1 1 0 0 1 1-1c2 0 4.5-1.2 6.24-2.72a1.17 1.17 0 0 1 1.52 0C14.51 3.81 17 5 19 5a1 1 0 0 1 1 1z\"\u002F>\u003Cpath d=\"m9 12l2 2l4-4\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":21},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"m22 7l-8.991 5.727a2 2 0 0 1-2.009 0L2 7\"\u002F>\u003Crect width=\"20\" height=\"16\" x=\"2\" y=\"4\" rx=\"2\"\u002F>\u003C\u002Fg>","\u003Cblockquote>\n\u003Cp>The slow part of posting long notes to Xiaohongshu is not “generation.” It is \u003Cstrong>judging pagination and tone on the cards themselves\u003C\u002Fstrong>. Do not let the AI finish a whole carousel in chat. Let it write faster into the page you have open; you decide which card is cramped and which line does not sound spoken — that is human–AI cowriting.\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Cp>\u003Cimg src=\"\u002Fblog\u002Fai-xiaohongshu-cards-from-notes\u002Fcover.webp\" alt=\"Core argument illustration\">\u003C\u002Fp>\n\u003Ch2>Why does pasting a note into Xiaohongshu always need a rewrite?\u003C\u002Fh2>\n\u003Cp>Xiaohongshu wants a \u003Cstrong>carousel of images\u003C\u002Fstrong>. A note wants a \u003Cstrong>long text stream\u003C\u002Fstrong>. Those two forms disagree about how long a paragraph can be. In Obsidian, a blog, or meeting notes, eight hundred words under one heading is normal. A vertical Xiaohongshu card usually holds a title plus a few bullets; more than that is cramped, less looks empty.\u003C\u002Fp>\n\u003Cp>So the failure is almost never “bad copy.” It is \u003Cstrong>not seeing the split before you post\u003C\u002Fstrong>. Ask an AI in chat to “make this Xiaohongshu-ready” and you often get still-too-long prose, or a set of images you cannot fix one at a time. Whether pagination works, and whether a card sounds spoken, has to be judged on the card.\u003C\u002Fp>\n\u003Cp>Technical notes add friction: code, jargon, and lists that work in a long article turn into a wall of type once they sit on a card. Regenerating the whole set from a new prompt often rewrites the cover card you already liked.\u003C\u002Fp>\n\u003Ch2>How are the cards split?\u003C\u002Fh2>\n\u003Cp>By default each \u003Cstrong>second-level heading (\u003Ccode>##\u003C\u002Fcode>)\u003C\u002Fstrong> becomes one card; insert \u003Ccode>---\u003C\u002Fcode> when you need a harder break. A note with six \u003Ccode>##\u003C\u002Fcode> headings usually becomes six images — merge by demoting \u003Ccode>##\u003C\u002Fcode> to \u003Ccode>###\u003C\u002Fcode>, split by breaking a section and inserting \u003Ccode>---\u003C\u002Fcode>.\u003C\u002Fp>\n\u003Cp>That rule is stabler than “let the model pick the page count,” because pagination follows structure you already wrote, not the model’s mood that day. Common fixes:\u003C\u002Fp>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>What you see\u003C\u002Fth>\n\u003Cth>Likely cause\u003C\u002Fth>\n\u003Cth>What to change in the source\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\n\u003Ctr>\n\u003Ctd>One card is too dense, no whitespace\u003C\u002Ftd>\n\u003Ctd>The paragraph under this \u003Ccode>##\u003C\u002Fcode> is too long\u003C\u002Ftd>\n\u003Ctd>Insert \u003Ccode>---\u003C\u002Fcode> in the middle, or turn a paragraph into a list\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Two cards that should be one\u003C\u002Ftd>\n\u003Ctd>The second \u003Ccode>##\u003C\u002Fcode> should not stand alone\u003C\u002Ftd>\n\u003Ctd>Change the later \u003Ccode>##\u003C\u002Fcode> to \u003Ccode>###\u003C\u002Fcode>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Thin cover, then too many fragments\u003C\u002Ftd>\n\u003Ctd>No short title card at the top\u003C\u002Ftd>\n\u003Ctd>Add a \u003Ccode>##\u003C\u002Fcode> that is only a title and one-line takeaway\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Tone reads like a paper, not spoken\u003C\u002Ftd>\n\u003Ctd>Long sentences, passive voice, stacked jargon\u003C\u002Ftd>\n\u003Ctd>Edit only that card’s paragraphs; do not rewrite the whole note\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\n\u003C\u002Ftable>\n\u003Cp>For knowledge notes, 3:4 or 4:5 vertical is common; lists and one-liners can go 1:1. Theme and typeface can wait — \u003Cstrong>get pagination right before you chase looks\u003C\u002Fstrong>, or a new skin will not save a card stuffed with three paragraphs.\u003C\u002Fp>\n\u003Ch2>How do I let my current AI edit the card page?\u003C\u002Fh2>\n\u003Cp>Have the assistant write into \u003Cstrong>the card editor you already have open\u003C\u002Fstrong>, not a second draft in chat, and not a finished image set you only review afterward. Open \u003Ca href=\"\u002Fmarkdown\u002Fxiaohongshu\u002F\">Markdown to Xiaohongshu cards\u003C\u002Fa>, paste the note, turn on Connect your AI, keep the tab open, then go back to Cursor, Claude, or whichever assistant you already use.\u003C\u002Fp>\n\u003Cp>One round looks like this — the AI speeds the split and the tone; you watch the preview and decide which card is good enough:\u003C\u002Fp>\n\u003Cp>\u003Cstrong>You:\u003C\u002Fstrong>\u003Cbr>\nSplit this Git note into cards by second-level headings. Card 2 should keep only three short bullets in a spoken tone. Leave the later cards alone for now.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Assistant:\u003C\u002Fstrong>\u003Cbr>\nAlready written to the card page you have open. The preview now shows 4 cards; card 2 is three bullets. Check length and tone on the cards — if one is off, name the number and I will only change that card.\u003C\u002Fp>\n\u003Cp>Stop there. Do not imagine what the page looks like — chat on the left, an arrow to the tool page on the right, preview already refreshed from that write:\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"\u002Fblog\u002Fai-xiaohongshu-cards-from-notes\u002F02-ai-flow.en.webp\" alt=\"Chat on the left, MeTool card page on the right showing the split preview\">\u003C\u002Fp>\n\u003Cp>If one card is cramped, do not say “start over.” Say “split card 4 from the second paragraph.” Theme and ratio can wait — \u003Cstrong>pagination first, looks second.\u003C\u002Fstrong> That is cowriting on this page: production speeds up on the assistant side; quality stays on a preview you can see. Concept: \u003Ca href=\"\u002Fblog\u002Fp\u002Fhuman-ai-cowriting\u002F\">What is human–AI cowriting?\u003C\u002Fa>. Setup: \u003Ca href=\"\u002Fblog\u002Fp\u002Fconnect-ai-to-online-tools\u002F\">How do I let my current AI operate a web tool?\u003C\u002Fa>. Hub: \u003Ca href=\"\u002Fai\u002F\">Let Your AI Use MeTool\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>On this page the assistant sees a virtual file \u003Ccode>main.md\u003C\u002Fcode>. You can also say:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>“Add one takeaway line on the cover card. Do not touch the later cards.”\u003C\u002Fli>\n\u003Cli>“Card 4 is cramped. Split it from the second paragraph.”\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>On export, ask the assistant to list options first: pack all cards, or download only one. Then trigger the browser download — image bytes do not come back through chat; the final filename is whatever the save dialog uses.\u003C\u002Fp>\n\u003Ch2>What is worth slicing into cards, and what is not?\u003C\u002Fh2>\n\u003Cp>Notes that \u003Cstrong>already have heading structure\u003C\u002Fstrong> are worth slicing. Unstructured dumps, or pieces that must keep long copy-paste code, are not.\u003C\u002Fp>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>What you have\u003C\u002Fth>\n\u003Cth>Better path\u003C\u002Fth>\n\u003Cth>Why\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\n\u003Ctr>\n\u003Ctd>Obsidian \u002F technical blog with clear \u003Ccode>##\u003C\u002Fcode>\u003C\u002Ftd>\n\u003Ctd>Slice first, then let AI fix one card\u003C\u002Ftd>\n\u003Ctd>Pagination matches the original structure\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>A wall of text with no headings\u003C\u002Ftd>\n\u003Ctd>Ask AI to add \u003Ccode>##\u003C\u002Fcode> first, then slice\u003C\u002Ftd>\n\u003Ctd>Without \u003Ccode>##\u003C\u002Fcode>, the default is a handful of oversized cards\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Meeting notes, SOP, step lists\u003C\u002Ftd>\n\u003Ctd>The docs-side knowledge-card flow\u003C\u002Ftd>\n\u003Ctd>More 3:5 \u002F 1:1 explainers; see \u003Ca href=\"\u002Fblog\u002Fp\u002Fai-markdown-to-xiaohongshu-cards\u002F\">How do I turn a Markdown doc into knowledge cards?\u003C\u002Fa>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Must show long copyable code\u003C\u002Ftd>\n\u003Ctd>Do not force it onto images\u003C\u002Ftd>\n\u003Ctd>Code on a card is hard to copy; readers bounce\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>One caption, no carousel\u003C\u002Ftd>\n\u003Ctd>Write it in Xiaohongshu directly\u003C\u002Ftd>\n\u003Ctd>The split cost is not worth it\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\n\u003C\u002Ftable>\n\u003Cp>Boundaries: the connection aims at this tab only; close it and the assistant has no destination. Cards render in the browser, for notes you are about to post — not for bulk-slicing someone else’s unpublished writing. AI can speed tone and breaks. It cannot tell you whether a card will be swiped past in the feed — that glance is still yours.\u003C\u002Fp>\n\u003Ch2>Wrap-up\u003C\u002Fh2>\n\u003Cp>Long notes on Xiaohongshu start from a form mismatch: a text stream has to become a carousel, split on \u003Ccode>##\u003C\u002Fcode> and \u003Ccode>---\u003C\u002Fcode>. Put the note on a card page with a live preview, let AI edit the same \u003Ccode>main.md\u003C\u002Fcode> faster, and you point at the card that is wrong. Do not let the model finish the whole set in chat — structure first, theme later. The middle you skip is usually the rework.\u003C\u002Fp>\n",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":24},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M15 14c.2-1 .7-1.7 1.5-2.5c1-.9 1.5-2.2 1.5-3.5A6 6 0 0 0 6 8c0 1 .2 2.2 1.5 3.5c.7.7 1.3 1.5 1.5 2.5m0 4h6m-5 4h4\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":26},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M15.707 21.293a1 1 0 0 1-1.414 0l-1.586-1.586a1 1 0 0 1 0-1.414l5.586-5.586a1 1 0 0 1 1.414 0l1.586 1.586a1 1 0 0 1 0 1.414z\"\u002F>\u003Cpath d=\"m18 13l-1.375-6.874a1 1 0 0 0-.746-.776L3.235 2.028a1 1 0 0 0-1.207 1.207L5.35 15.879a1 1 0 0 0 .776.746L13 18M2.3 2.3l7.286 7.286\"\u002F>\u003Ccircle cx=\"11\" cy=\"11\" r=\"2\"\u002F>\u003C\u002Fg>",1788171087508]