Open the explorer. Load the synthetic sample, acknowledge the recorded-example notice, and choose “View recorded check-in.” Values are transcribed from the video: overall 79; acne 82; redness 99; moisture 72; texture 79; pore 65.
Choose Comfort & hydration. Moisture 72 appears as context. It does not select or rank products.
Set budget ₹1,500, mark Cleanser as owned, and choose Lightweight. Keep fragrance-free checked. Expect gel ₹449 plus sunscreen ₹699: ₹1,148, with ₹352 remaining.
Reduce budget to ₹1,100. Expect a ₹48 gap and disabled Add. Restore ₹1,500.
Add only missing steps. The bag should contain two products. Remove the gel: subtotal ₹699. Close and re-add the missing step: two products totaling ₹1,148, without duplicate sunscreen.
Download the decision card; clear the recorded check-in. The planner remains usable without image signals.
What is demonstrated
A shopper-controlled budget/reuse decision with explicit cost reasoning. Seven catalog entries are illustrative. No checkout, clinical suitability claim, or external user-study result is presented.
Live Node build
Requires Node 20 or later. Clone the source, copy .env.example to .env, and configure YOUCAM_API_KEY and the assigned India API origin privately. Run npm start. The secret stays on that server. The repository’s README explains the signed-upload, task, polling, and privacy flow.