Publish Xiaohongshu notes and watch comments on a schedule
A product manager in China spent a month testing OpenClaw with an open-source Xiaohongshu MCP server to publish notes, reply to comments and collect competitor data. The setup worked, but logins and the connector broke often.
What they did
He ran the open-source xiaohongshu-mcp project as a bridge between OpenClaw and the browser, logged in by scanning a QR code and passing the session cookie to OpenClaw. He then had OpenClaw publish notes, reply to comments, search and collect note data for competitor analysis and try attaching product links. For comment handling he used a cron schedule plus a library of approved replies.
What happened
The poster says publishing, comment replies and data collection all worked through the agent, and he estimates running costs stay low for a small team. He shares no reach or follower numbers.
Keep in mind
Cookies expire every one to two weeks, the MCP server often stopped and needed a fresh QR login or restart, and earlier scripted approaches triggered captchas and risk controls. Live-stream interaction and bulk direct messages were not supported, and unofficial automation carries account risk.
Try it with your agent
Using the Xiaohongshu connector, search notes about <your product category> and summarize the titles, tags and comment themes of the top <number> by likes. Draft one note for <your product> in that style and save it as private or draft for my review. Every <time interval>, check new comments on my notes, draft replies from <approved replies file>, and leave anything about pricing, complaints or orders for me.
Replace the <angle brackets> with your own details.