How do I export Xiaohongshu comments in bulk to find what users complain about? (2026)
Search notes on the topic, take each note's comments page by page, and give the table to your AI to sort by complaint. One page of search and 3 pages of comments under each of 20 notes are 61 calls at $0.0200: $1.22, for about 600 comments. DataXAPI reads Xiaohongshu (RedNote) with 21 endpoints, from $0.0200 a call; a failed call is not charged.
Updated 2026-11-04
How to do it
With an AI that has DataXAPI connected, say: “Search Xiaohongshu for this product, take 3 pages of comments under the top 20 notes, and list the complaints: what, how many times, and a line in the user's own words.”
Without an AI: xiaohongshu.notes.search gives the notes; xiaohongshu.notes.comments.list takes a note's ID or link and gives about 10 comments a page, newest first: text, commenter, likes, replies, time. Use dataxapi call with --pages and --out to write them to a file. The dataxapi run comments Skill writes a CSV, but a run is capped at $0.05, which is 2 pages here.
Good to know
These are top-level comments. Replies are a separate call, xiaohongshu.comments.replies.list, about 5 a page at $0.0200: take them only under comments that matter. Sort by likes first: a complaint many people agreed with counts for more than one that was repeated once.
Commenters' names are personal information: leave them out of what you analyse and share.
Endpoints
Each with its price; every page has the parameters, the fields and a recorded response.