Use Cases

Dotient makes searching your visual files instant and intuitive. Here's how different people use it in practice.

The Storm Chaser Building a Tornado Taxonomy

A storm chaser has 4,000 photos from a decade of chasing, everything dumped into folders by date. She wants to find every wedge tornado versus every rope tornado without having gone through and tagged them at the time. She creates two Signals: feeds one 15 photos of wedge tornados as positives and a mix of clear sky and rope tornados as negatives, does the same for rope tornados. Now the embedding space is reshaped around her own visual vocabulary, and she can type “wedge tornado at sunset” and get compositional results that combine her trained concept with a natural language query the base model was never taught to recognize. Signals aren't a tag, they're a reshaping of how the whole library gets searched from that point on.

The Bird Photographer Separating Species After a Shoot

She gets home from a wetland reserve with 800 photos spanning a dozen species, all in one Import folder because sorting on location is a waste of daylight. She drags the whole batch into a Cubby, hits the cluster button, and watches Dotient auto-group them by visual similarity, herons in one cluster, egrets in another, a stray group of blurry flight shots off to the side. She manually drags three misclassified juveniles into the right cluster, since young birds don't always cluster clean, then renames each group and moves on to editing. What used to be an hour of squinting at thumbnails is five minutes of drag and confirm.

The Architect Digging a Floor Plan Detail Out of a 200-Page PDF Set

He remembers a specific stair detail drawing but not which of the eleven PDF binders it's in, and definitely not the page number. He searches “spiral stair section” and Dotient returns the extracted image from inside the PDF directly, not the whole document, so he opens straight to the page instead of scrubbing through a 200-page file.

The E-Commerce Seller Finding Every White-Background Product Shot

She's shooting products for a shop and has years of photos mixed with personal photos on the same drive. Rather than a folder discipline she never kept, she runs “product photo on white background” as a text query since it's a visual concept the base SigLIP model already understands well without any training.

The Researcher Pulling a Graph Out of a Stack of Papers

Mid-thesis, she remembers a scatter plot showing a specific correlation but has forgotten which of 60 downloaded papers it came from. Dotient's PDF handling surfaced the extracted image at import time, so a search for “scatter plot with two clusters” finds the figure and she opens directly to the source paper.

The Interior Designer Building a Mood Board Live With a Client

Sitting with a client, she pulls up Cubbies and starts dragging in reference shots from a shared drive while they talk, using the cluster button to group by palette in real time. Instead of pre-building three separate boards and hoping one lands, she builds it live and the client watches shots visually group into “warm neutral” and “high contrast” clusters as they're dropped in.

The Freelance Illustrator Finding “That One Texture” for a Commission

He has thousands of scanned paper textures and grain overlays and remembers a rough one he used two years ago but never named. A visual “find similar” search off a thumbnail he vaguely recognizes surfaces the exact file plus a dozen close variants he'd forgotten he had.