Visual similarity / duplicate image detection

Near-identical, re-encoded, cropped and lightly edited copies are found across a large library, not just byte-for-byte duplicates. It is used both to clean archives before processing and to check whether an image has appeared somewhere before. Similarity thresholds are yours to set, so what counts as a duplicate matches your policy rather than a fixed default.

You can use the models we operate or bring your own model and we run it on your cluster. Work arrives as an API call, a batch over a folder or bucket, or a resident on-premise service, and results return as structured output with a per-job audit record.