The most powerful image search on Hive — and it gets sharper every single day.
Most search is basically reading over your shoulder. When you look for a photo, it isn't looking at photos at all — it's scanning the words underneath them. The title. The tags. The caption the photographer happened to type. If nobody wrote "golden hour," you'll never find that golden-hour shot, no matter how perfectly the light falls.
Shape doesn't work like that. Shape looks at the picture.
Every photo published on liketu gets read by Shape — not the words around it, the actual image. The light. The color. The composition. The mood. Who or what is in the frame, and how it's framed. Shape turns all of that into something it can search, so when you go looking, you're searching the photography itself — not the description someone may or may not have bothered to write.
That one difference is why Shape is the most powerful image search we've ever put in your hands. Here's what it actually feels like to use.
Type what you're picturing. Describe the shot in your head — "misty mountains at sunrise," "a quiet street café in the rain," "bold red against concrete" — and Shape finds photos that look like that. Not photos that contain those words. Photos that match the feeling. The photographer never had to use your language; Shape understood the image on its own.
Search with a photo. Have an image you love? Drop it in. Shape finds its closest visual cousins across the whole community — same energy, same palette, same composition — in about a second.
Follow a creator's eye. Love how someone shoots? Shape can surface work that feels like theirs, even from creators you've never heard of. It's the fastest way we know to fall down a beautiful rabbit hole.
Then refine. See something close? Say "more like this, but warmer," "but no people," "at night" — and Shape adjusts. It feels less like querying a database and more like talking to someone with very good taste.
When Shape reads a photo, it doesn't slap a few keywords on it. It converts the image into a kind of mathematical fingerprint — a long list of coordinates that capture what the photo is actually about. Two photos that feel alike end up with fingerprints that sit close together. Two that feel nothing alike sit far apart. It's a map of meaning, and every photo on liketu has a place on it.
Here's the clever part: your words live on the same map. When you type "golden hour on a misty ridge," Shape drops that sentence onto the very same map the photos live on — then simply looks at what's nearby. So a phrase and a picture can be compared directly, even though one is made of language and one is made of light. That's why Shape isn't thrown off by a missing caption, a wrong tag, or a different language. It isn't matching letters. It's matching meaning.
That's the whole leap from old search to Shape: keyword search asks "did someone write this word?" Shape asks "does this photo mean what you mean?"
Here's the part we love most. A search like this gets more accurate as the community gets bigger — the exact opposite of feeling crowded.
Think back to the map. Every new photo published is a new point on it. The more photos there are, the denser the map gets — and the denser the map, the closer the nearest match is to whatever you're picturing. With a few thousand photos, the closest thing to your idea might be "pretty close." With millions, there's almost certainly a photo sitting right on top of the exact vibe in your head. Every creator who joins, every shot that goes up, doesn't just add content — it sharpens every search that comes after it.
So Shape isn't a feature that ships and stalls. It compounds. The liketu you search next year will be measurably better at finding your photo than the one you search today — with zero effort from you. The community does the work just by being itself.
That's the idea. Stop searching captions. Start searching photographs.
🔗 liketu.social — explore our features
🔍 liketu.social/shape — Shape, search by image
📲 liketu.social/download — get the android app
🏃 liketu.social/health — StepJar, Health Loop step pools
🗺️ liketu.social/atlas — Atlas
❓ liketu.social/faq — questions, answered
image-first. mobile-first. built on Hive.
