Xiaohongshu RedNote Note to Text

Turn a Xiaohongshu (小红书 / RedNote) note into text an LLM can read. Image notes too: every photo is described and its text read out. Optional viral breakdown (爆款拆解): the hook, the beats image by image, and a reusable template.

Official website: linkdigest.dev — powered by LinkDigest, which also offers this as an MCP server and a REST API.

The problem this solves

Most of Xiaohongshu is 图文 — a caption plus a stack of photos. yt-dlp only understands video notes, so the majority of the platform returns nothing. The words that matter are inside the images.

This Actor returns the post instead — every image described and OCR'd, the caption, and the key points:

{
  "platform": "xiaohongshu",
  "author": "周星星",
  "title": "从 Agent Team 到 Agent Swarm（下篇）",
  "transcript_source": "none"
}

Input

Field	Type	Notes
urls	array of strings	Required. xhslink.com/… or xiaohongshu.com/discovery/item/… — share links work as-is.
format	json or markdown	json gives separate fields; markdown gives one ready-to-paste document per link.
breakdown	boolean	Optional viral breakdown (爆款拆解), one extra credit per link — see below.
insight	boolean	Optional consumer insight (消费者洞察): audience, claims, objections, purchase-intent cues, products and prices, questions readers ask — every item a quote checked against the post and its top comments. One extra credit per link; arrives as an insight field.
maxItems	integer	Optional ceiling so a long list cannot cost more than expected.

This Actor only reads xiaohongshu links. Paste anything else and it says so rather than failing quietly. For every platform in one place — Xiaohongshu, Douyin, TikTok, YouTube, X and ordinary web pages — use apify.com/loongnian714/social-url-to-llm-text.

Viral breakdown (爆款拆解)

Tick breakdown to have each post taken apart the way a content strategist would: the hook in its first seconds (the opening words copied exactly), the structure as timed beats, the title formula, the call to action, who it is for, and a reusable template with placeholders. Every quote is checked word for word against the post and removed if it is not there; likes, comments, saves and hashtags come from the platform, not the model. It adds one credit per link and arrives as a breakdown field (and a section of the Markdown).

Output

One dataset item per link: platform, author, title, posted_at, caption, transcript ({t, text}), ocr_text, images ({description, ocr}), key_points, source_url, transcript_source, degraded, cached, credits, and with breakdown on, breakdown.

A link that cannot be read still produces an item carrying url and error. Nothing is silently dropped — a short dataset is the worst way to find out that three of your fifty links failed.

Pricing

Event	Price
digest-completed	$0.02 per link successfully read
extra-credit	$0.01 per credit beyond the first

Credits track the two things that actually cost money: images to describe and minutes of media to transcribe. One base credit covers a typical post and its first six images; each further six images adds one, each started minute of audio or video adds one, and a viral breakdown adds one.

A link that fails is never charged. Neither is a run that fails because of a problem on our side.

Reading competitor notes at volume (对标账号)

Brands, agencies and MCNs use this to turn a folder of competitor and benchmark notes into one table: caption, the text on every image, the transcript, key points and the viral breakdown. Paste the whole list into urls. Every link gets its own dataset item, so the run exports straight to CSV or Excel, and maxItems caps the spend.

If your team reads hundreds of notes a week, the same reader is cheaper used directly from linkdigest.dev: a batch endpoint takes 50 links per call, posts a signed webhook when the batch is done, and is reachable over REST and MCP. Teams paying by card can subscribe to the Studio plan ($49 a month for 3,500 credits). Teams in mainland China can buy a ¥288 pack of 2,000 credits with Alipay; it is not a subscription and never expires. The workflow is described in Chinese at linkdigest.dev/zh/duibiao-zhanghao.

Notes

• Already-digested links return in about a second — anything anyone has run before is cached, and cached links cost one credit, never more.
• The heavy work runs on our servers, not on Apify: reading these platforms takes session cookies, ffmpeg, speech recognition and a vision model.
• Chinese text is returned as-is, not translated.