How do I find complaints about a brand on Xiaohongshu?

Complaints on Xiaohongshu rarely say "complaint". They look like a note titled "避雷！这款面霜千万别买" with 300 comments agreeing, or a calm review where the real problem sits in the replies. A brand team that only searches its brand name sees mostly ads and fan posts. The trick is to search the way unhappy users write.

Step 1: search with complaint words

Run Xiaohongshu note search for the brand or product name combined with words like 避雷 (steer clear), 踩雷 (got burned), 翻车 (went wrong), 拔草 (talked out of buying), 难用 (hard to use), 过敏 (allergic), 售后 (after-sales). Use sort=latest for monitoring, sort=popular for the notes doing the most damage.

Step 2: read the comments where it matters

For the notes with the highest comment counts, pull note comments. Comments carry text, likes, reply counts and IP location; sorting by likes shows which complaints other users agree with.

Step 3: filter

A keyword filter (refund, fake, broke, rash, customer service) catches most of it. An LLM pass grouping comments into themes catches the rest. Keep the note link next to each quote so someone can reply.

Cost of a weekly report

Step	Calls	Cost
5 complaint phrasings × 2 pages	10	$0.20
Top 10 notes × 3 comment pages	30	$0.60
Per week	40	{{cost:xiaohongshu.notes.search10+xiaohongshu.notes.comments.list30}}

Example

for w in 避雷 踩雷 翻车 售后; do
  curl -s -G "https://api.okdata.online/v1/xiaohongshu/notes/search" \
    -H "Authorization: Bearer $OKDATA_API_KEY" \
    --data-urlencode "query=某品牌 $w" -d sort=latest \
    -d fields=id,title,url,metrics.comments,published_at
done

Limits

• Top-level comments only. Each carries its reply count; heated threads show up as comments with many replies.
• Search matching is fuzzy; read before escalating.
• Only public content. Complaints in private groups or messages are not visible to anyone outside them.
• Every page is a separate charge, so cap pages per phrase.

Alternatives

Social listening suites do this with dashboards and alerting and suit teams that want no scripts. MediaCrawler can collect the same notes and comments with a logged-in account. For the full routine, including Douyin and Weibo, see word-of-mouth monitoring.