- Templated pages Templated pages at scale MonZap
A price guide generated per Pokémon, set and mechanic.
Proof: 41 Espeon entries on one page, 89 Eevee entries on another, the same FAQ skeleton on each ("How much is a X card worth?"), TCGplayer and Cardmarket reference prices, and a "Download MonZap Free" button on every page. Source: monzaptcg.com/pokemon/espeon/ and /pokemon/eevee/ · read 2026-10-25
Open
- Confident wrong answer Wrong card, priced as certain MonZap
A wrong identity, served with a price.
Proof: Two photos of the same card, Erika's Victreebel 006/217, answered as "Sprigatito, Scarlet & Violet Black Star Promos 76" and priced at $12.81. Photo 84's answer carried the run's "locked" state. Source: Run bench100, photos 24 and 84 · 2026-10-24 (run page and CSV)
Open
- Quota burn Failed scans eat your quota Center Stage
A wrong scan costs the user one scan from the quota.
Proof: App Store review, verbatim: "When it scans the cards it will show you a different card and not give you back the scan so when it goes from 500 to 499 and it was the wrong card that shows up it takes that scan from you and now you have less scans left due to their own mistakes." Source: App Store review · 2026-09-29
Open · vendor announced a change; retest queued
- Unsourced accuracy Its own listicle ranks it first CardGrader.AI
No lab test behind the ranking was found.
Proof: Its "Best Pokémon Card Scanner Apps in 2026" guide ranks its own app first. Source: cardgrader.ai guide · documented 2026-10-21
Open
- Unsourced accuracy Invents the rivals' numbers
Other apps' scores, stated as findings, with no run behind them.
Proof: "Independent testing puts Ludex at about 85% and CollX at about 65%." No test is published. The same site cites a 10,000-card benchmark study that does not exist in public. Source: pregradecards.com · read 2026-10-21
Open
- Unsourced accuracy 98.7% match rate, no method
No sample, method or dataset behind the number.
Proof: "98.7% match rate with professional human grading", sold on its blog. No sample, no method, no dataset. The same blog admits version 1 was shut down for inconsistent results. Source: Vendor blog · read 2026-10-03
Open
- Unsourced accuracy 95% alignment, no count
The claim and the guide that repeats it are the same source.
Proof: "95% alignment with PSA grades" on its site, with no count, sample or method. The claim sits over its own "Best Pokémon Card Grading Apps 2026" guide. Source: cardmintai.com · read 2026-10-04
Open
- Unsourced accuracy 92.8% by press release CardGrade.io
A launch number, not a measurement. The two inspection counts cannot both be right.
Proof: "92.8% accuracy, ±0.22 grade variance", announced by press release. Its own materials disagree on whether it inspects 47 points or 16 zones. Source: Launch press release · read 2026-10-03
Open
- Contradicted claim Markets ~99%; its own test says 96% on 20 cards
The marketing number and the measured number are both the vendor's. They disagree.
Proof: "Roughly 99% accuracy" on its product pages. Its own published test reports 96% on 20 cards, with rivals partly unscored. Source: eyevo product pages and published test · read 2026-10-21
Open
- Star ratings, no counts Stars instead of counts
No counts are published, so the ranking cannot be checked.
Proof: 15 apps tested on 30 cards, accuracy collapsed into stars, no raw counts published. Source: bindex test · documented 2026-10-21
Open
- Clone A ChatGPT wrapper sold as a grading service
Nothing is measured. Your photo goes to ChatGPT.
Proof: A ChatGPT wrapper on yeschat.ai, sold as a grading service. No photo is measured by the vendor. Source: yeschat.ai · read 2026-10-03
Open
- Clone Centering calculators in the grader list
A centering tool is one pillar of four. Listed as graders, they read as the whole measurement.
Proof: Two of three centering-calculator clones in the same store list, presented among grading apps. Source: App Store search · read 2026-10-03
Open