UPI Testing
Save your workspace key — shown once.

8F8K-56CB-SAYV-8ZVG

Why it matters
It is the only way to reach these profiles and results from another device. The lab stores a hash of it, so nobody here can recover it for you.

PhonePe

PH-003 Not applicable here

PhonePe — With merchant category code

No mc set, so this is the same payload as the plain control.
Why
Profile P03 is defined by mc, and neither your UPI profile nor this run supplies one — so comparing P03 against P01 would be comparing a method against itself and reading the noise as a finding. Set mc on the profile first, or run P01 and know what you measured.
pa — A payment needs a payee address. Nothing can be launched without pa.
You are on a desktop browser, which has no UPI app to open. The URI is still shown and can be copied.
What would unblock it
Open this page on the phone you want to test, or use a QR-scan method (QR group) which is designed for exactly this situation.
Test PH-003

Scan with the phone you want to test.

Advanced details
phonepe://pay?pa=&tr=LABPGW6SH2KN9JX&am=2.00&cu=INR
Launch
app scheme
Scheme
intent:// (Android) · phonepe://pay (iOS)
Profile
P03 — P01 + mc
Platform
Android + iOS
Browser
Android: Chrome or any browser that follows intent: URIs. iOS: Safari, only where the app registers a scheme.
Docs
No public third-party deep-link specification. PhonePe publishes a merchant SDK/PG integration instead. (observed)
Reference
LABPGW6SH2KN9JX
Registry
1.0.0

THE most important comparison in the catalogue. Several apps switch from a person-to-person flow to a merchant flow when mc appears — changing limits, fees and whether the amount can be edited. P01 against P03 on the same app, same destination, same minute is the controlled experiment that shows it.

Expected

PhonePe opens on its payment confirmation screen showing the payee and ₹2.00. THE most important comparison in the catalogue. Several apps switch from a person-to-person flow to a merchant flow when mc appears — changing limits, fees and whether the amount can be edited. P01 against P03 on the same app, same destination, same minute is the controlled experiment that shows it.

Limitations

  • A browser cannot tell you whether an app is installed. "App did not open" and "app is not installed" look identical from here; AN-004 in the Android harness is what separates them.
  • The iOS scheme used here is observed, not vendor-documented. A working result is evidence, not a supported integration.

Prerequisites

  • A saved UPI profile with a payee address.
  • On iOS: PhonePe installed and registered for phonepe://pay
  • On Android: PhonePe installed (com.phonepe.app).

Notes

  • This scheme is observed, not vendor-documented. A result either way is evidence about today, not a supported integration.

How the lab runs it

Android: an intent:// URI naming com.phonepe.app, with no browser_fallback_url so a failure to open stays visible. iOS: the phonepe://pay scheme carrying the same query.