Xiaohongshu Creator Notes Scraper: RedNote User Posts

Extract recent public Xiaohongshu (RedNote / XHS / 小红书) creator-note cards from known profile pages. Export titles, covers, creator identity, pinned status, source-exposed engagement counts, and note IDs/links when available, with up to 30 cards requested per profile.

This Actor uses direct HTTP to read the profile page's server-rendered data. No Xiaohongshu login or login cookies are supplied by the user; it does not start a browser or configure a proxy. This is public-page extraction, not access to private or login-only posts. Apify account/API authentication is still required to run it through Apify.

What is supported

There is one creator-profile mode. It reads the first published-note group embedded in each profile response; it does not request subsequent pages, even if the response contains a cursor or hasMore flag. The 30-note setting is a ceiling, not a complete-history promise.

Useful applications include creator content sampling, comparing public note-card engagement, and building datasets from known creators. It does not fetch full note bodies, comment threads, video files, or every post the creator has ever published. Pinned cards and source ordering mean position is not an independently verified chronological rank.

Quick start

Enter creator references in the Input tab. This illustrative input uses the schema's prefilled profile; accessibility must be checked in your own permitted run.

{
  "profileUrls": [
    "https://www.xiaohongshu.com/user/profile/65c381340000000011014d37"
  ],
  "maxNotesPerProfile": 20,
  "minimumNotesPerProfile": 1,
  "failOnProfileError": true,
  "maxConcurrency": 3,
  "requestTimeoutSecs": 15,
  "maxRequestRetries": 2
}

Accepted references include /user/profile/<id> URLs, xhslink.com profile-share links, copied share text containing a URL, and 24-character hexadecimal user IDs. Use profile shares rather than individual-note shares.

Input fields

Field	Runtime default	Behavior
profileUrls	No runtime default	Required array; schema accepts 1–100 references. Known IDs and identical normalized URLs are deduplicated.
maxNotesPerProfile	20	Integer, 1–30, upper bound per profile. The Input UI prefill is 3, not the omitted-field runtime default.
minimumNotesPerProfile	1	Integer, 1–30, no greater than the maximum. Profiles below this extraction threshold are recorded as errors.
failOnProfileError	true	Any invalid or failed selected input fails the run before dataset rows are written. With false, failed profiles are skipped and successful profiles may be saved.
maxConcurrency	3	Integer, 1–10 concurrent profiles.
requestTimeoutSecs	15	Integer, 5–40 seconds per HTTP request.
maxRequestRetries	2	Integer, 0–3 retries after the first attempt.

A legacy userIds array is recognized when profileUrls is absent. New integrations should use profileUrls. proxyConfiguration is ignored by the runtime; there is no supported proxy mode. No downstream Actor is called.

Output and billing unit

Each saved dataset row is one extracted note card, available through Apify dataset exports or its API. Fields can include:

• recordId, input, requestedUrl, profileUrl, userId, nickname, and one-based source position.
• Optional noteId, noteUrl, title, type, and isPinned.
• coverUrl, optional coverWidth/coverHeight, source, and scrapedAt (UTC ISO extraction time).
• likedCount, likedCountText, likedCountLowerBound, likedCountIsExact, plus source-exposed collectedCount, commentCount, shareCount and their text fields.

If a stable note ID is absent, recordId is a deterministic hash of creator/ID/title/cover values. It is not a guaranteed permanent note identifier: changes to title or cover can change the hash. A noteUrl may be constructed when an ID exists, but tokens can be missing or expire; a link is not a promise of future access.

Exact numeric engagement fields are omitted for abbreviated source values rather than invented. For likes, inspect the separate text, lower-bound, and exactness fields. Exactness refers to the parsed source representation, not independent measurement. Comment and share counts do not include comment text or replies. Missing fields remain missing.

Result thresholds and limits

Results are collected before any dataset write. With failOnProfileError: true, an invalid input, blocked profile, or profile below the configured minimum causes a pre-write failure; diagnostics go to the ERRORS key-value-store record. With false, partial output is allowed, but a complete run with zero valid note rows still fails.

The minimum is an extraction check, not a promise of delivery after budgets or output caps. In particular:

• Apify Free plan: only the first 50 normalized valid creator inputs are processed, and the concatenated output is then truncated to 50 dataset rows across the run. Later profiles can have no saved rows even after passing their extraction threshold.
• Paid Apify plans: no developer-set 50-row cap; the 100-reference schema limit, per-profile 30-note ceiling, runtime limits, and event-charge budgets still apply.
• Profile references are deduplicated before crawling, but there is no global note-row deduplication after extraction. Unresolved alias shares can produce overlapping rows.
• Private, deleted, restricted, empty, or blocked profiles may provide no cards. Markup and access-policy changes can break extraction. There is no guaranteed speed or complete history.

Current pricing

Prices are USD per event; Free-plan results are not free.

Event / tier	Price
apify-default-dataset-item — FREE	$0.014 per saved note row
Same event — BRONZE	$0.0028 per saved note row
Same event — SILVER / GOLD / PLATINUM / DIAMOND	$0.0025 per saved note row
apify-actor-start — all tiers	$0.0005 per startup event

Startup events are charged per GB of allocated memory, minimum one event. At the default 256 MB, one minimum startup event applies; do not assume a fixed fee for every memory setting. Startup can be charged even when the run returns no rows.

The runtime uses Actor.push_data for the final dataset write and no extra custom per-note charge. ERRORS records do not generate result events. Apify platform usage is separately paid by the running user; compute, transfer, and storage can still cost money on empty or failed runs. Event fees are not total run costs. Confirm the current tier-specific rates on the Pricing tab.

FAQs

Can I export a creator's full posting history? No. Only the first published-card group exposed in the retrieved profile response is used, capped at the requested maximum.

Does the minimum guarantee apply when errors are allowed? No. With failOnProfileError: false, profiles that miss the extraction threshold are skipped. The Free-plan global cap and charge budgets can also reduce saved output.

Why are IDs, links, or numeric counts missing? The profile response may omit IDs/tokens or show abbreviated engagement. The Actor preserves available card data instead of inventing unavailable details.

Collect and retain public personal data only for a lawful purpose, consistent with applicable laws and platform terms.

Complementary Actors

• Xiaohongshu Profile Scraper — creator bios and public profile counters.
• Xiaohongshu Influencer Finder — discover creator candidates from category note cards.