小红书 cards: turn long content into vertical illustrations

小红书 (Xiaohongshu) consumes content as stacks of vertical cards, not paragraphs. If you have a long-form post (a blog, a research summary, a tutorial) and you want it on 小红书, the right move is not “post a screenshot of the article” — it’s generate purpose-built cards.

Step 1: Generate the card series

Use document-illustrator by op7418. Feed it your long-form source and ask for 3:4 vertical format with gradient-glass style — this combination photographs well in the 小红书 preview grid.

Use document-illustrator on my AI product launch blog post.
Generate 6 vertical 3:4 illustrations + a cover.
Style: gradient-glass.

Step 2: Polish the cover with Canvas Design

The first card determines whether someone scrolls. Sometimes document-illustrator’s auto-cover is fine; sometimes you want editorial typography that the auto-generator doesn’t deliver. Use canvas-design for a hand-tuned hero card with a single bold serif headline.

Use canvas-design to make a 3:4 vertical hero card. Title:
"AI 产品发布的 3 个反直觉决定". Editorial style, warm cream
background, single accent color (warm orange).

Step 3: Write the platform-native caption

A 小红书 post has a long caption + hashtags below the cards. Use wechat-article-writer — its “personal hands-on review” or “identity transformation” voice maps to 小红书 norms (more first-person, more sensory detail, less claim-and-prove).

Use wechat-article-writer to write a 小红书 caption for my AI launch
post. Style: personal hands-on review. Length: 200-300 chars.
Add 5-8 relevant hashtags at the end.

Output checklist

For one 小红书 post, you ship:

6-8 cards (Step 1) + a hero (Step 2), all 3:4 vertical
1 caption (Step 3) of ~200 chars
5-8 hashtags
(Optional) 1-2 emoji to anchor the caption opening

Variations

Tutorial post — use document-illustrator’s “vector illustration” style instead of gradient-glass. Reads more diagrammatic.
Product review — fewer cards (4-5), more text-heavy hero.
Cross-post to 微信公众号 — re-run the pipeline with 16:9 horizontal output for the WeChat editor instead. See the WeChat article scene.

What this pipeline doesn’t do

Short video — for video shorts use youtube-clipper to clip + caption. (小红书’s video format is its own animal.)
Live photography — none of these generate real photos. You’ll need to shoot or stock-source separately.

But for the bread-and-butter 小红书 post — long content adapted to vertical card format with platform-native voice — this stack runs in ~15 minutes.