RedNote (Xiaohongshu) Trending & Category Feed Scraper

Need more than the trending feed? Xiaohongshu (RedNote) All-in-One Scraper adds search, comments, creators, topics and products in the same run, with no login. Your channel and fullDetail carry over. Feed notes cost $3 per 1,000 there.

Discover what's trending on 小红书 right now — no keyword, no login, no cookie. Pick one of 12 categories or the main recommended feed and get fresh notes with engagement data as clean JSON rows.

This actor reads the same category feeds the logged-out web app renders — fashion, food, beauty, movies, career, love, home, gaming, travel, fitness, video, or the main mixed feed — and each reload rotates in fresh notes, so pulling more pages keeps yielding new content. When Xiaohongshu won't render a category for a logged-out visitor (frequent since September 2026), the rows come from keyword search on the category's name instead, tagged origin: "keyword-search".

What you get per note

Field	Description
noteId, noteType	ID and type (normal image post or video)
noteTitle	Display title
likedCount / likedCountNum	Likes as shown (e.g. 1.2万) and parsed to a number (12000)
userId, userName, userAvatar	Author
noteCoverUrl	Cover image
channel	Which category feed produced the row
noteUrl + xsecToken	Ready to paste into the Note Detail or Comments scraper

Flip on fullDetail and every discovered note is hydrated to its complete record — description, full image gallery, video URL, all counts — at one extra request per note.

Example input

{
    "channel": "homefeed.cosmetics_v3",
    "maxPages": 3,
    "maxItems": 50
}

FAQ

Really no login? Really. No cookie or account is involved. For a specific keyword rather than a category, use the Search Scraper, which also runs without login.

Why do repeated runs return different notes? The feed rotates on every load — that's the point. It's a sampling instrument for what the platform is pushing right now, not a fixed index. Duplicates within one run are skipped automatically; the run stops early after three reloads with nothing new.

How long does a run take? Usually 20-60 seconds. If you set a run timeout, give it 120 seconds or more. With a shorter one the run stops a few seconds before the timeout, keeps the rows it has found, and says so in its status message, rather than being cut off by the platform.

What's this good for? Trend monitoring per vertical (what beauty content is surfacing today), creator discovery without keywords, and building seed lists for the Note Detail and Comments scrapers.

Related scrapers

• RedNote Search Scraper — keyword search, no login
• RedNote Note Detail Scraper — hydrate any discovered note, no login
• Xiaohongshu (RedNote) Scraper — the all-in-one actor: 8 operations

🤖 For AI Agents & LLM Apps

Compact reference for AI agents calling this actor via the Apify MCP server or the Apify API (actor: memo23/rednote-trending-scraper).

Purpose: Samples Xiaohongshu's (RedNote / 小红书) logged-out discover feed — 12 categories or the main recommended feed — returning fresh trending notes as JSON rows. No login.

Minimal input:

{ "channel": "homefeed_recommend", "maxPages": 2, "maxItems": 40 }

Output: one row per note; key fields: noteId, noteType, noteTitle, likedCount, likedCountNum, userId, userName, userAvatar, noteCoverUrl, channel, noteUrl, xsecToken. With fullDetail: true, rows carry the complete note record (description, imageUrls, videoUrl, all counts).

Behaviors an agent should know:

• The feed rotates per load — runs are samples, not a stable index; rows dedup within a run and the run stops after 3 reloads with nothing new.
• channel enum: homefeed_recommend (default) plus homefeed.{fashion,food,cosmetics,movie_and_tv,career,love,household_product,gaming,travel,fitness,video}_v3.
• fullDetail: true costs one extra request per note.
• A run usually takes 20-60 s; set the run timeout to 120 s or more. Under a shorter timeout the run finishes a few seconds early with the rows found so far (status SUCCEEDED, status message says it stopped early), so a short timeout can return fewer rows than maxItems, or none.
• No cookie, no login. Billing: pay-per-event — see the Pricing tab on the actor page.

⚠️ Disclaimer

This actor collects publicly displayed data from Xiaohongshu for legitimate research, marketing, and archival purposes. You are responsible for complying with applicable laws and Xiaohongshu's terms. Do not use collected data for spam or harassment.