What ComicCheck checks first: your private collection
ComicCheck starts with the records that belong to the collector: owned issues, quantities, ratings, wishlist choices, collecting goals, and price preferences. Those records are used to answer the buying question; they are not published on this methodology page. When the device has a usable local cache, an exact duplicate check and basic one-issue gap check can still run without waiting for an external catalog.
An imported comic does not disappear just because an external service cannot identify it. ComicCheck preserves the normalized series title, issue number, series start year when supplied, variant notes, barcode, and other available fields. That minimal record can still prevent a duplicate and can participate in conservative numerical adjacency checks. A local match is evidence about what the collector owns, not proof that every edition detail is known.
The table above describes reviewed product behavior, not a live uptime or availability report. A source marked active can still be temporarily unreachable, return no matching record, or provide incomplete fields. ComicCheck should preserve the useful local answer in those cases and label the missing evidence instead of inventing it.
How title and issue search works
The active public catalog path uses a bounded, same-origin, cached title-and-issue lookup backed by the Grand Comics Database. ComicCheck normalizes ordinary title punctuation and issue-number formatting, then checks the returned candidates rather than treating the first result as certain. The lookup is bounded to protect reliability and provider capacity. It is not exhaustive, and a missing result does not establish that a comic does not exist.
The optional series start year identifies the beginning of a publishing run. It is not the issue's cover date or publication year. Adding it can separate two runs with the same title and issue number. When that field is omitted, ComicCheck may use a recent-release check to surface a current run, but that is a retrieval heuristic rather than a complete historical search.
A title-and-issue match can still leave printing, cover designation, direct-market versus newsstand edition, or another variant attribute unresolved. When more than one plausible edition remains, the result requires confirmation before ComicCheck presents an exact-edition duplicate warning. A broad issue-level match may still support a qualified recommendation when the user has chosen to treat different variants as the same issue.
Barcode and cover-photo limits
A UPC can narrow identification, but it does not always encode enough information to prove an exact printing or cover. The current public GCD path is not a complete barcode index, so barcode lookup can return several candidates or no candidate. ComicCheck must show the candidate it thinks it found and preserve an ambiguity warning rather than silently upgrading an issue-level match to an exact-variant match.
The PWA cover-photo beta uses Tesseract.js to extract title and issue hints in the browser. It is not visual-similarity matching, grading, or authentication. The current OCR path does not upload or retain the selected image. Extracted text can be wrong because of logos, stylized lettering, glare, bags, price stickers, or crowded cover design, so the collector must confirm the suggested comic.
Future image-similarity recognition would be a separate candidate-retrieval step with its own provider, rights review, retention controls, and confidence language. It would still require confirmation whenever the evidence cannot distinguish the base issue from a particular cover or printing.
Metadata rights are not cover-image rights
The GCD site identifies its database material under Creative Commons Attribution-ShareAlike 4.0, and ComicCheck provides source attribution when that bibliographic metadata is used. Metadata rights do not automatically grant rights to cover art. Cover scans can remain copyrighted even when issue facts are reusable under an open-data license.
ComicCheck therefore treats provider cover scans as identification media, not as a blanket-cleared image library. The production path does not display or retain GCD cover scans without verified permission for that use. A future image from Wikimedia would also need a file-level review: Wikidata licensing explains the structured-data terms, while Wikimedia Commons reuse guidance makes clear that reusers must evaluate the information attached to each file. Neither source creates automatic clearance for every comic cover.
Metron's documentation is useful for evaluating a possible future connector, but Metron is not active in the production lookup described here. Production use remains permission-gated until commercial use, caching, redistribution, attribution, and cover-image terms are confirmed in writing. Comic Vine is excluded from the production source plan under the current licensing and commercial-use posture.
How data quality changes recommendations
ComicCheck separates recommendation strength from data confidence. Collection fit asks how well the candidate connects to the collector's runs, story interests, creators, characters, wishlist, and other stated goals. The buy decision then applies the asking price and hard rules such as an exact duplicate. Data confidence reports how much reliable identity and relationship evidence was available for those conclusions.
Recommendation explanations come from the rules that ran. An exact locally cached duplicate can support a high-confidence warning even while the network is unavailable. In contrast, missing variant, storyline, event, creator, or reading-order information should reduce confidence or remove a related reason. It must not be replaced with a guessed relationship or a mysterious score.
Provider records are normalized into ComicCheck's internal comic identity rather than copied as one large user-owned object. This lets local records survive provider changes and helps prevent a provider outage from erasing the collector's own evidence. It does not make incomplete provider data complete.
Corrections, source changes, and review date
This page is reviewed when a source, license interpretation, retention behavior, or shipped identification path materially changes. The visible date and methodology version identify that review; they do not certify that every remote service is healthy at the moment a visitor reads the page. Operational failures should be reported in the product as limited availability, not hidden by changing this methodology description.
Collectors can use Wrong Comic? in the buying flow when a candidate is incorrect. That correction is about the collector's current decision and does not publish their collection. ComicCheck does not currently publish a verified rights-reporting address, so third-party cover display remains disabled until a monitored process exists. This page will be updated before that policy changes.