How Says Hi tools work

How a reading is produced

Open a tool’s site and offer a photo, a name, or a word. A photo is resized in your browser first, then sent — with your optional note — to that tool’s own service, which passes it to Qomza’s AI service. An AI model returns a structured reading, and the tool presents it with its uncertainty labeled.

This directory runs none of that itself: it has no upload, no model, and no data store. It only introduces the tools and links to them.

A clearer photo can help a reading — when a result comes back uncertain, a sharper shot is worth trying, though uncertainty can also reflect the tool’s own limits. Photo tips that work across every Says Hi tool covers the cross-subject fundamentals, and each product’s own photo guide carries the specifics.

Readings are AI, plainly

Every reading is generated by an AI model. A model can be wrong — about a breed, a coin, a plant, or a word’s root — so a reading is a starting point for your own look, not a verdict, a certificate, or a diagnosis.

How uncertainty is shown

The tools name suggestions rather than settle them: matches and cues are labeled “possible” or “likely”, and when a photo leaves details unclear the reading says so instead of guessing confidently. The exact wording varies by tool — each site’s own how-it-works or methodology page explains its labels — but the contract is shared: a labeled suggestion, not a measured certainty.

What the tools don’t do

A photo can’t confirm an identity, a health problem, an allergen, an authenticity, or a value — so no Says Hi tool offers medical, veterinary, safety, appraisal, or authentication answers. Where a question needs a professional — a vet, a doctor, a mechanic, a grading service — the product’s own site points that way.

How these pages are made

The pages on this directory are written and maintained by the operator. Each product site publishes its own guides, and they follow a shared practice: a guide names the sources it drew on and links them, notes when those sources were last checked, and credits or labels its images — generated illustrations are marked as such. Each product’s own methodology page describes the details for its guides.

Where a page carries a “Last updated” date, it marks the last substantive change to that page. This directory’s own imagery, icons, and fonts are credited on the Credits page.

Corrections

Reports are read, and a page is corrected when an error is confirmed; the page’s “Last updated” date changes when that happens. To flag a wrong reading or a problem with this directory, see Report a problem.