Verification is the product
A QR platform that decode-tests every code before you print it, gets payment QR correct in every wallet, and never lets a printed code die.
The problem we started from
A QR code is the only piece of design nobody can proofread. You cannot look at it and tell whether it works, whether the contrast is sufficient, whether the logo has eaten too much of it, or whether it will still read at 18mm on a table tent under a warm lamp. So people print five hundred of them and find out afterwards.
The second problem is worse, because it is deliberate. Across this category, a code stops resolving when a plan limit is hit, a subscription lapses, or a trial ends. The printed object is fine; the redirect behind it has been switched off. Menus, packaging and signage go dead for a billing reason.
What we built instead
Every code is scored against eighteen checks and decode-tested at five resolutions under eight degradations before you download it — live, next to the preview, never blocking you. You get a measured minimum print size in millimetres, and the specific fix for anything that fails.
Payment QR gets the same treatment. A UPI, PIX, QRIS or EPC URI is validated against its specification and then against how each wallet actually parses it, because Google Pay rejects what PhonePe silently repairs. The compatibility matrix never shows a pass for something that fails in a real app.
And a printed code never stops resolving. Limits degrade measurement, not function. If you cancel, your codes enter Continuity Mode and keep redirecting to their last destination, permanently.
Why we publish our measurements
Claiming to predict scan reliability puts an obligation on us to show our work. Before writing product code we spent six weeks building a calibration corpus of twelve thousand codes, decoding each with four independent decoders, then physically printing four hundred stratified samples and scanning them with six devices under three lighting conditions.
The dataset and method are published, including the parts that were inconvenient. Our print-size predictions land within ±15% of measured failure thresholds, and where we are uncertain we say so rather than rounding it into a marketing number.
How this is built
One person, working with AI tooling, on a deliberately thin architecture. That shapes the product in ways worth being open about: opinionated defaults over deep configurability, no features that need a human in the loop, and every module behind a kill switch so a misbehaving feature can be turned off without a deploy.
It also means support scales through documentation rather than headcount, and that we would rather ship four excellent surfaces than twenty mediocre ones. Where we have chosen parity rather than innovation — barcodes, short links, integrations — we say so.
Non-negotiables
These are architectural, not aspirational. Each one has a regression test that runs on every release, and each is expensive enough to reverse that stating it publicly is the point.
- 01Tested every release
A redirect never waits on analytics, billing or AI
Scan events go edge → queue → store, never synchronously and never through the app database on the hot path. Killing the analytics pipeline does not affect a single redirect.
- 02Tested every release
A printed code never stops resolving
Not on downgrade, cancellation or limit breach. Limits degrade measurement, never function. This is enforced in the data model, not just promised in copy.
- 03Tested every release
No advertising in the scanner journey, ever
The person scanning your code is not our customer and must never be monetised. No ads, no interstitial upsells, no watermark — on any tier, with no exception path.
- 04Tested every release
AI never generates or modifies an encoded payload
AI may propose form values that you confirm. The payload is always deterministic and always from confirmed input, because a hallucinated VPA sends someone’s money to a stranger.
Honest by default
No countdown timers, no exit-intent modals, no fake urgency, and no logo wall of companies who are not customers. Limits are stated in plain language before you hit them.
We have no customer testimonials on this site. Not because nobody uses the product, but because we will not publish invented ones — we will show customer numbers once we have real ones worth showing.
Our business model, stated plainly
We make money from subscriptions, AI credits, API access and white-label licensing. We do not sell scan data, aggregated intent signals, or any behavioural data product — and there is no plan under which that changes.
Subscriptions
Five paid tiers, monthly or annual. You pay for volume, automation and collaboration — never for correctness.
AI credits
Metered usage for generation and extraction features. Hard stop at zero, never a surprise charge.
API as a product
Scanability scoring and payment validation are sold standalone to print shops, packaging houses and fintechs.
White label
Agencies and resellers licence the platform under their own brand, with per-client billing attribution.
Never for sale: scan data, aggregated intent signals, or any behavioural data product.
Built in India, correct everywhere
Data for Indian customers is held in ap-south-1 under DPDP 2023, with a named grievance officer published in our privacy notice. We are a URI generator and print-asset producer — not a payment aggregator, PSP or TPAP.
Who runs this, and who reviews what we publish
qrcodegenerator.in is operated by Agochar Tech LLP, a limited liability partnership registered in India, from 415 Decora Square, Sadhu Vasvani Road, Rajkot 360005, Gujarat, India.
The pages that tell you whether a code will work — the payment pages, the problem pages and the blog — are reviewed by the qrcodegenerator.in team at Agochar Tech LLP. Each one names the team and the date it last changed.
If something on one of those pages is wrong, tell us which page and what you saw. Corrections are made on the page itself, and its date moves. Write to [email protected].