My Xiaohongshu scraper keeps getting blocked: what does keeping it up cost against paying per call? (2026)
30 keywords searched once a day come to $18 a month at $0.0200 a page, with no account, cookie or proxy of yours. A scraper of your own has no fee per call: what it costs is accounts, proxies, a server and the hours of repair each time it is blocked. Add those up for a month and divide by the price, and you have the number of calls where the two meet. DataXAPI reads Xiaohongshu (RedNote) with 21 endpoints, from $0.0200 a call; a failed call is not charged.
Updated 2026-11-04
What each side pays for
Most of a scraper's cost is not on any invoice, which is why it is easy to count as free.
| Item | Your own scraper | Per call |
|---|---|---|
| Accounts | Yours, and a blocked one has to be replaced | None of yours: one key |
| Proxies and a server | Yours to buy and run | None |
| Hours of repair | Yours, each time a request is stopped or the site changes | None: you get an error, and a failed call is not charged |
| Fee per call | None | $0.0200 a page of search |
| What you can get | Whatever you write the code for | What the endpoint gives: no read counts on Xiaohongshu, for example |
Work out your own figure
A month of your own scraper = hours of upkeep × what an hour costs you + accounts + proxies + the server. Divide that by the price of a call.
An example with made-up inputs, to be replaced by yours: 10 hours of upkeep a month at $30 an hour is $300, before accounts and proxies. At $0.0200 a page, $300 is 15,000 pages, about 300,000 notes.
Calling less than that a month, paying per call costs less. Calling far more, with upkeep that stays small, your own can cost less. If the job has a deadline, count what a lost day costs you too.
What a month comes to per call
A page of search usually holds 20 notes with their numbers, so most jobs do not need a call per note. On the command line: dataxapi call xiaohongshu.notes.search query=咖啡 sort=latest --pages 5 --max-spend 0.1 --out notes.json.
| Job | Calls | Comes to |
|---|---|---|
| 10 keywords, one page a day | 300 × $0.0200 | $6 |
| 30 keywords, one page a day | 900 × $0.0200 | $18 |
| The first page of comments under 300 notes | 300 × $0.0200 | $6 |
| The whole text of 100 notes | 100 × $0.0200 | $2 |
Stay or switch
Keep your own if you want a few hundred notes once, for study; if you need something the endpoints do not give; or if someone already keeps it running with little effort. Switch if it runs on a schedule, sits inside a product, or someone is waiting on the results.
Before switching, know what search leaves out: text is the result card's excerpt, and there are no read counts or follower counts. xiaohongshu.notes.get has the whole text and xiaohongshu.users.get the followers, at $0.0200 a call. It reads public data only and cannot post. How to keep a scraper from being blocked is not something this page covers.
Endpoints
Each with its price; every page has the parameters, the fields and a recorded response.