How we verify every code.
Most coupon sites show you codes and hope they work. We run a three-layer verification engine that tests every code before you see it. When we make a claim, we show the evidence behind it. That's the Glass Box.
Most coupon codes on the internet are junk.
Coupon sites scrape codes from social media, forums, email leaks, and affiliate feeds. They list everything they find. No testing. No verification. No accountability.
The result: you copy a code, paste it at checkout, and watch it fail. Then you try the next one. And the next. Ten minutes later you've tried eight codes and none of them worked.
This is not a minor inconvenience. It's a broken system. Coupon sites profit from showing you content regardless of whether it works. Their incentive is volume, not accuracy.
Three layers. Independent verification.
No single method is enough. Automated testing catches most failures. Human verifiers catch edge cases machines miss. Fleet signal detects changes in real time. Truth emerges when independent layers converge.
Automated Code Testing
Headless browsers simulate real checkout flows across our merchant network every day. They add items to cart, navigate to checkout, enter codes, and observe the result. No guessing - actual testing on actual sites.
This is the foundation layer. It produces binary signals at scale: did the code change the cart total, or not?
merchant: "nike.com"
test_code: "SAVE20"
cart_before: $119.99
cart_after: $95.99
delta: "-$24.00 (20.0%)"
result: FUNCTIONAL
adapter_tier: "PLATFORM_ADAPTER"
tested_at: "2026-02-12T08:23:00Z"
merchant: "chewy.com"
test_code: "FIRST30"
result: CONDITIONAL
condition: "new_customer_only"
adapter_tier: "HEURISTIC"
Human Verification Network
Tens of thousands of trained verifiers manually test codes and submit proof. Every submission includes a screenshot, a timestamp, and a verification of the discount applied.
Humans catch what machines miss: geo-restrictions, membership requirements, stacking rules, minimum order thresholds, and exclusion lists. This layer provides nuance that no automated test can replicate.
Real-Time Fleet Signal
Millions of SimplyCodes users generate real-time checkout signals. When someone uses our extension and a code succeeds or fails at checkout, that outcome feeds back into our verification engine.
This is the freshest layer. A code can pass automated testing in the morning and fail by noon because the merchant changed their terms. Fleet signal catches that within minutes.
Truth emerges when independent layers converge.
No single layer is authoritative. A code verified by automation but contradicted by fleet signal gets flagged. A code reported by one human but unseen by automation stays unverified. Verdicts require convergence across layers.
When there are no codes, we tell you.
Most coupon sites will never show an empty page. They pad results with expired codes, guesses, and "try this!" suggestions to keep you clicking. We do the opposite.
When our verification engine determines that no working codes exist for a merchant, we say so. Clearly. Definitively. A confident "no" is more valuable than a hopeful "maybe" - it saves you time and builds the trust that makes our "yes" mean something.
Stores like Apple, Costco, and Lululemon rarely run coupon promotions. We verify that reality instead of pretending otherwise.
A single number. Three inputs.
Every code gets a Health Score from 0 to 100. It combines outcome data, consensus weight, and freshness decay into one confidence measure. The higher the score, the more confident we are the code will work for you right now.
Health Scores decay over time. A code verified 2 hours ago scores higher than one verified 2 days ago, even if both passed. Freshness matters because merchants change their promotions constantly.
Every verdict comes with a receipt.
We don't ask you to trust us. We show you the evidence behind every verdict. This is the Proof Packet - a structured evidence chain documenting exactly how we reached our conclusion.
"code": "SAVE20",
"merchant": "nike.com",
"verdict": "VERIFIED",
"health_score": 96,
"discount": "20% off sitewide",
"evidence": {
"automated_test": {
"timestamp": "2026-02-12T08:23:00Z",
"result": "FUNCTIONAL",
"cart_delta": "-$24.00",
"adapter_tier": "PLATFORM_ADAPTER"
},
"human_consensus": {
"verifications": 4,
"positive": 3,
"avg_trust_score": 0.92,
"confidence": "HIGH"
},
"fleet_signal": {
"sample_size": 47,
"success_rate": 0.94,
"trend": "STABLE",
"last_success": "2m ago"
}
},
"conditions": {
"min_order": 50.00,
"exclusions": ["sale items"],
"stackable": false
}
}
Read the full research.
Anatomy of a Verdict
The complete technical specification of our verification engine. How layers interact, how consensus is weighted, and how verdicts are synthesized.
Read the whitepaperLive Verification Feed
Watch the verification engine work in real time. See codes being tested, verdicts being issued, and Health Scores updating across 631K+ stores.
Watch liveSee it in action.
Install the extension and watch Health Scores, Proof Packets, and real-time verification happen at your next checkout.