Verify an Amazon AI image tag before upload
A file passes the AssetDisclose check only when the exact contains-synthetic-performer keyword is found in XMP dc:subject. A filename match, misspelling, wrong metadata field, or duplicate copies do not count as a clean result.
What a valid result means
A valid result confirms the required string is readable from the expected XMP field in the file being checked. It does not decide whether the visual content falls inside the policy or guarantee a marketplace decision.
Common verification failures
- The keyword is absent.
- The value is misspelled or uses spaces instead of hyphens.
- The keyword was written to a different metadata field.
- Multiple exact copies were appended.
- The application saved changes to a sidecar file rather than the upload file.
Recommended pre-upload workflow
Verify the exact file that will be uploaded, not the source document or an earlier export. If another app compresses, resizes, or converts the image after tagging, run the verifier again on the final output.
Evidence behind this guide
Policy guidance can change. Confirm the current Seller Central instruction before a production upload.
Common questions
Can metadata disappear after export?
Yes. Some image optimizers, design tools, and marketplace export workflows strip metadata, so verify the final deliverable.
Is one duplicate harmless?
A duplicate may still be readable, but it is unnecessary and creates inconsistent metadata. AssetDisclose normalizes the exact value to one copy.
Does verification upload my image?
No. The current verifier reads supported image files locally in your browser.
Amazon AI media tagging resources
Move from policy interpretation to writing, verification, batch processing, and provenance.