Photograph or scan a card front and back. The analyzer measures centering geometrically, profiles the four edges and corners for wear, scans the surface for scratches, print lines and spots, then maps the result onto PSA / BGS / SGC / CGC standards. Centering is objective; everything else is an estimate and is reported with a confidence band. Every grade this tool outputs is an estimate, not a guarantee of what a grading company will assign.
| Best | Flatbed scanner, 1200 dpi, card face-down on glass, lid closed, no sleeve. Gives dead-flat geometry and repeatable light. Centering error < 0.5%. |
| Good | Copy stand or tripod directly overhead, two diffuse lights at 45°, card on a matte background that contrasts with the border (dark grey for white-bordered, light grey for black-bordered). |
| Phone | Camera parallel to card, card filling ~70% of frame, indirect window light, no flash. Perspective error is the main accuracy killer — the analyzer corrects it, but tilt beyond ~15° degrades the centering measurement. |
| Surface pass | Scratches and gloss breaks on holo/chrome are invisible in flat overhead light. Take a second shot with a single light raking across the card at 20–30° to reveal them. |
| Never | Shoot through a penny sleeve, toploader or slab. Reflections read as surface defects and the sleeve edge breaks card detection. |
Ranges show the plausible spread given image confidence, not a guarantee. Graders apply human judgment, eye appeal and (at PSA) qualifiers that no image pipeline reproduces.
The centering engine is deterministic. The wear detectors are thresholded heuristics — tune them against cards whose real grades you already know, then export the profile so runs stay comparable.