Xiaohongshu (RedNote) Content API — Trending & Notes

Pay per note. Pull Xiaohongshu (RedNote / 小红书) trending posts, topical category feeds, and full note detail as clean JSON — no login, no session cookie, no Mandarin-speaking analyst required.

Feed a mode and get back structured note records: the cookieless recommend feed, a topical category channel (fashion, food, cosmetics, travel and more), or full detail for specific note URLs. It is a keyless, code-friendly window into China's earliest publicly observable consumer-demand signal — where nearly every viral note is a product review or buying guide, surfacing weeks before Tmall or JD sales rankings.

What this is, in one paragraph

Xiaohongshu (小红书, "Little Red Book", known outside China as RedNote) is where Chinese consumers discover and review products. This actor turns its public discovery surfaces into flat JSON. In trending mode it reads the recommend feed; in category mode it reads a specific topical channel; in note mode it fetches full detail (title, body text, tags, and engagement counts) for note URLs you supply. Trending, category, and note modes all run cookieless — no login and no session token. Every record carries the note's title, creator display name, engagement, media links, and provenance. No private contact data of any kind is collected, and any email, phone number, or messaging handle found in note text is stripped before output.

Who buys this and for what job

• Brand and sourcing teams at Western consumer companies reading China's product-demand signal before it reaches sales rankings.
• Trend and market researchers tracking what is going viral by category — beauty, fashion, food, travel — over time.
• Content and social teams benchmarking creators and note formats that win engagement on RedNote.
• Builders and AI agents who need a keyless RedNote content endpoint callable from code.

The job it does: read what China is talking about and buying, by topic, as structured data — without an account, a scraper to maintain, or a Mandarin-speaking analyst in Shanghai.

Pricing

Pay per event. A small fixed fee when a run starts, then one event per note delivered; the note is delivered before it is charged. Tiered by your Apify plan. No subscription.

Event	Free plan	Bronze	Silver	Gold / Platinum / Diamond
Actor start	$0.05	$0.05	$0.05	$0.05
RedNote note	$0.02	$0.018	$0.016	$0.0134

About 50 notes per dollar at the base note rate. The default limit of 25 keeps evaluation cheap; raise it for full runs.

Input

Field	Type	Default	Description
mode	select	trending	trending (recommend feed), category (a channel feed), note (detail by URL), or search (keyword/hashtag — needs a cookie).
category	select	recommend	Used when mode = category: fashion, food, cosmetics, movie, career, emotion, home, gaming, travel, fitness.
note_urls	array	[]	Used when mode = note: full note URLs (with xsec_token) to fetch in detail.
keywords / hashtags / users	array	[]	Used when mode = search. Search is login-gated by Xiaohongshu.
cookie	string (secret)	—	A logged-in web cookie. Only needed to unlock search; trending, category, and note modes run without it.
limit	integer	25	Maximum notes delivered and billed.
proxyConfiguration	object	Residential	Residential proxy is strongly recommended for reliable access.

Output

One JSON object per note. Real example (trending):

{
  "record_id": "6a5eeb0d000000001301abcd",
  "note_id": "6a5eeb0d000000001301abcd",
  "mode": "trending",
  "category": null,
  "title": "我的乡村生活 户外美食",
  "description": null,
  "note_type": "video",
  "author": "户外明泽",
  "author_id": "5f0e...",
  "liked_count": 3998,
  "collected_count": null,
  "comment_count": null,
  "share_count": null,
  "tag_list": null,
  "has_video": true,
  "cover_url": "https://sns-webpic.../c.jpg",
  "note_url": "https://www.xiaohongshu.com/explore/6a5eeb0d...?xsec_token=...",
  "xsec_token": "...",
  "source": "Xiaohongshu (RedNote)",
  "source_url": "https://www.xiaohongshu.com/explore/6a5eeb0d...",
  "observed_at": "2026-08-25T00:00:00Z"
}

In note mode the record additionally carries description, tag_list, collected_count, comment_count, and share_count. An unbilled RUN_RECEIPT records the source check, the mode, how many notes were delivered, and any partial notices.

Field reference

Each record carries record_id and note_id, the mode and category that produced it, the note's title and (in note mode) description — both with any private contact string removed — note_type, the public creator author and author_id, engagement (liked_count, and in note mode collected_count, comment_count, share_count), tag_list, has_video, cover_url, and the canonical note_url with its xsec_token. Provenance travels on every record: source, source_url, licence, attribution, disclaimer, and observed_at.

How a run works

For trending and category modes the actor reads Xiaohongshu's server-rendered discovery feed and normalises each note card; because the recommend feed rotates, it pages the surface until your limit is reached, de-duplicating by note id. For note mode it fetches each note's page and extracts the full detail. Search mode targets the keyword/hashtag surface, which Xiaohongshu gates behind a logged-in request — supply a valid cookie to use it. Every note's text is passed through a contact-data scrub (email, phone, and messaging-handle patterns) before the record is written, and the platform bills one note event per delivered record.

Working with the data

Records are flat and keyed by note_id, so they load straight into a warehouse or a notebook. Use category and tag_list to segment by topic, liked_count and the note-mode engagement counts to rank what is resonating, and has_video to split formats. Run the same category on a schedule to build a time series of what is trending in a product space. Feed a batch of note_urls to enrich notes you have already found elsewhere with full detail and engagement.

Use with AI agents

Point Claude, an OpenAI Agents SDK tool, an n8n or Make flow, or any MCP-aware client at this actor. Because the output is flat JSON, an agent can pull a category feed, rank notes by engagement, and summarise the emerging themes in one pass. A typical prompt: pull the top trending cosmetics notes on RedNote and summarise the products people are excited about.

Why cookieless matters

Most RedNote scrapers demand a logged-in session cookie for everything, which ties your pipeline to one account that can be rate-limited, flagged, or logged out at the worst moment — and forces you to rotate credentials to run at any scale. This actor keeps its three core modes — trending, category, and note detail — fully cookieless, reading Xiaohongshu's own server-rendered discovery pages the way a first-time visitor's browser does. That means no account to maintain, no credential to leak, and no single point of failure across a large run. A cookie is only ever needed for the one surface Xiaohongshu genuinely gates — keyword search — and even then it is optional and clearly marked, so you always know exactly which part of a run depends on it.

Honest limitations and caveats

• Search is login-gated. Trending, category, and note modes are cookieless and reliable. Keyword/hashtag/user search requires a valid logged-in cookie; without one, no search results render and the run says so.
• Feeds rotate. The recommend and category feeds are personalised and time-varying; two runs return different, overlapping sets rather than a fixed ranked list.
• Public content only — no private contact data. Only public note content and the creator's public display name and id are returned. No email, phone, or messaging handle is collected, and any such string found in note text is removed before output.
• Use a residential proxy. Xiaohongshu restricts datacenter traffic; the default residential proxy is strongly recommended for reliable access.
• Respect creators' rights. Note text and media belong to their creators and to Xiaohongshu; this data is for research and trend analysis, not republication.

---

Public content from Xiaohongshu (RedNote / 小红书) discovery surfaces, reformatted as structured records. Note text and media belong to their respective creators and to Xiaohongshu; returned for research and trend-analysis use. No private contact data of any kind is collected or emitted.