The workflow

How CardRevive AI Grading Works

The workflow separates capture, image preparation, focused condition analysis, and report assembly so you can see what evidence went in and what kind of output came back.

From upload to estimate

  1. Upload the evidence

    Choose a front image, a back image, or both. The service technically requires at least one image; both sides give the analysis more visible evidence.

  2. Review the crop and rotation

    The browser attempts to find and straighten the card. You can adjust the crop and rotation before confirming what will be graded.

  3. Validate and process

    The server checks the file signature and size, then prepares full-card and diagnostic views used to examine visible condition.

  4. Analyse four condition areas

    Focused analysis separates centering geometry from visible corner, edge, and surface observations. Unsupported evidence remains unavailable instead of being presented as measured fact.

    Deterministic visible-evidence example

    Corner, edge and surface observations stay separate

    Select a numbered marker to connect one synthetic visible cue to its own cautious observation.

    Generic card with corner marker selectedSynthetic whitening marker on the lower-right corner.

    Visual-only evidence · marker 1

    Corner

    Synthetic whitening marker on the lower-right corner.

    Measured
    Reserved for repeatable geometry that can support a ratio.
    Visual only
    Corner, edge and surface cues are described without invented dimensions.
    Unavailable
    Anything outside this synthetic front view is not assessed.

    These are synthetic visible markers, not authentication or an in-hand inspection. Illustrative pre-grading estimate; not an official grade.

  5. Assemble the report

    The service returns findings, company-style estimate ranges, likely outcomes, confidence notes, and recommendations. One completed assessment normally uses one grading credit.

  6. Keep, delete, or share

    The report and stored image renditions appear in your private library. You can delete a card or explicitly enable a public report, then revoke that public access later.

Before you spend a credit

Check image quality

Plastic, glare, shadows, clipped edges, blur, and aggressive filters can hide or imitate defects.

Read the capture guide.

Check card coverage

Standard rectangular cards are the clearest use case. Unusual shapes, finishes, and sizes can reduce measurable evidence.

Review supported cards.

Privacy through the workflow

Uploaded card images are processed for grading, sent to OpenAI's API for the focused AI analysis described in the Privacy Policy, and stored as resized WebP renditions with the report in your account library. Reports are private by default because public sharing consent defaults to off.

CardRevive does not use identifiable uploaded card images to train a separate CardRevive model. The Terms permit use of anonymised, aggregated grading data—not personal information or identifiable card images—to improve the service. Read the full Privacy Policy before uploading a sensitive image.

Your privacy choices

Essential sign-in, security, payments and on-device grading always work.

Optional first-party analytics, Google Analytics and Ads, Meta Pixel, plus Sentry error and session replay stay off unless you accept.

Read the Cookie Policy and Privacy Policy.