Guest mode, strictly
No login, no likes, no comments, no reposts — pure observation with zero account risk.
A daily research agent that scrolls Weibo's recommendation feed item by item, captures every video with 1M+ views, and returns a clean Markdown table — views, title, account, publish time, link.
No login, no likes, no comments, no reposts — pure observation with zero account risk.
View counts read from the app's UI dump, with screenshot OCR as backup.
Copying links tags the device, so the recommendation feed surfaces more of what you're hunting.
The agent opens the Weibo app and enters the recommendation feed anonymously — no login, ever. It scrolls one item at a time without skipping, reading each video's view count from the UI (with screenshot OCR as fallback). Anything under 1 million views is passed without stopping; matches get their link copied via the share menu, which also tags the device so the recommendation feed gets sharper every run.
Each hit is logged immediately — number, views, title, account, publish time, link — and the run continues until 20 matches are collected or 100 items have been scrolled, ending with a Markdown table summary. Strict constraints keep it clean: no likes, comments, follows, or reposts; nothing published; login popups closed and ignored.
A 20-row Markdown table of 1M+ view videos lands every day — scroll work done for you.
A week of tables reveals which hooks, lengths, and topics keep crossing 1M.
Each run's link-copies train the recommendation algorithm toward your research target.
Open Airtap, choose Weibo Viral Content Researcher, and it arrives pre-built with the whole flow ready.
Your accounts, budgets, thresholds, and non-negotiables - in plain language, no code.
Run it once, daily, weekly, or monthly - on a cloud phone or your own device.
Airtap watches outcomes, escalates when blocked, and reports back when the goal lands.