Published August 13, 20262 min
What it does
Post discovery agents scan the channels you choose every day, match new posts against rules you define — who you're looking for, which topics, what to avoid — and deliver a daily list of posts worth replying to. Every match comes with the reason it matched and the context you need to write a real reply.
Daily scan across channels
Xiaohongshu, Reddit, and X covered in one pass — including the Chinese-language channel most competitors never work.
Every match explained
No black-box scoring. Each post arrives with the reason it matched your rules, so you can trust or correct the judgment.
Feedback that compounds
Mark matches as useful or off-target. The next day's list gets sharper instead of noisier.
How it works
Define the rules once
Describe your target user, the topics and channels that matter, and your red lines — communities you won't touch, phrasing that signals a bad fit.
Agents scan daily
Each day, agents search the channels for new posts that match: recommendation requests, complaints about alternatives, questions your product answers.
You get a triaged list
Matched posts land in one list with the match reason, a summary, and a link. Fifteen minutes over coffee replaces an hour of tab-hopping.
Your feedback tunes the next run
Useful or not useful — one click per item. The rules sharpen over time, and the judgment stops living only in your head.
Why the reply window matters
Recommendation threads are usually decided within hours. A tool that surfaces posts weekly gives you dead threads; a person who checks twice a week misses the window entirely. Daily discovery with same-day delivery is what makes a reply land while the asker is still deciding.
FAQ
Which channels are supported?
Xiaohongshu, Reddit, and X today. The rules you define carry across channels, so adding a channel doesn't mean starting over.
How is this different from keyword alerts?
Keyword alerts match strings; discovery agents match intent. A post can contain your keyword and be worthless, or describe your exact use case without ever naming the category. Matches are judged against your rules and explained, not just pattern-matched.