Xiaohongshu Viral Post Teardown
Paste a Xiaohongshu image note: the caption, stats and in-image text are extracted automatically, then AI grades the sample, decodes cover elements, slide flow and factors, and outputs a one-variable test plan.
AI uses Qwen2.5 1.5B locally in your browser and requires WebGPU. First use downloads and caches about 1 GB; later runs reuse it. Clearing site data requires another download.
Mechanisms, not copies
Sample grading
Checks whether the numbers come from the creator or ads first; weak samples stop early.
Image text
Automatically reads the text inside every image from cover headlines to pages 2–5.
User path
From cover to purchase: which step is strong or weak.
14 factors
At least 8 factors, with the top 3 marked.
Variable layers
Core, flexible, careful and noise variables.
Controlled tests
A one-variable test plan with metrics and stop rules.
How to use
- 1
Paste the link
Paste the Xiaohongshu image-note link.
- 2
Add context
Add account type, followers, data, ads and your goal.
- 3
Get the teardown
After image text is extracted, AI writes the full teardown and test plan.
Who it is for
Xiaohongshu ops
Turn hits into testable hypotheses.
Seeding teams
Find the variables that really matter.
Creators & buyers
Understand why a hit worked and what to try next.
AI image teams
Write variant prompts variable by variable.
More free tools
Download & extract
By platform
FAQ
Numbers propped up by fame, ads or campaigns lead to false conclusions, so D-grade samples stop early.
It analyzes the extracted image text alongside your notes; purely visual details like lighting or color are flagged for checking against the original images.
Yes — A-grade samples with click or sales data are the most useful.
Image notes only; use Xiaohongshu Video for video notes.
Filtering out ad-boosted or follower-driven outliers helps focus your tests on repeatable content factors.