Use case
An e-commerce operator, reseller or archivist receiving a batch of new images and text needs to sort them into their existing category sheet for listing, filing or retrieval.
Sorting by hand against a spreadsheet, or feeding material to a generic model and fixing labels manually, or writing custom scripts against an API.
Generic recognition models only know fixed labels that do not match the user's own taxonomy, so items must be sorted by hand, which is slow and inconsistent at volume; the product page explicitly sells 'train AI on your own categories', indicating this mismatch is the pain it targets.
xOcto's call
Demand is evidenced
Trend: the taxonomy itself becomes a trainable asset instead of users adapting to a generic model's fixed labels. Entry: start from e-commerce catalogs, resale, archives and collections that already keep their own category sheets, and charge per volume sorted or per model maintained rather than selling a generic classifier.
Reason to use it
Why users would choose it
Inference: compared with manually relabeling each item, uploading one's own labeled samples and having the model output the matching category removes the step of correcting a generic model's labels one by one, so operators or archivists who already keep a taxonomy and handle large batches would choose it for bulk sorting; public material gives no accuracy or review method, so reliability remains unverified.
Where the easy answer breaks down
The tension worth following
An English validation note will follow from the public evidence.
If this is your job
Worth trying. Inference: compared with manually relabeling each item, uploading one's own labeled samples and having the model output the matching category removes the step of correcting a generic model's labels one by one, so operators or archivists who already keep a taxonomy and handle large batches would choose it for bulk sorting; public material gives no accuracy or review method, so reliability remains unverified.
Entry and what to borrow
Trend: the taxonomy itself becomes a trainable asset instead of users adapting to a generic model's fixed labels. Entry: start from e-commerce catalogs, resale, archives and collections that already keep their own category sheets, and charge per volume sorted or per model maintained rather than selling a generic classifier.