What Is This? Point Your Camera and Find Out

How do I find out what something is from a photo?

Fact: Lens App is a free download on the App Store with optional in-app purchases. The site gives you free daily scans on mobile and one free web scan per day, with paid tiers for people who scan often.

Why "what is this" is such an awkward thing to search

Text search assumes you already have words. Half the time you do not. A beetle is sitting on the screen door with orange spots and long antennae. A coin turned up in a drawer with worn lettering and a face you do not recognize. A plant in the rental yard has fuzzy leaves and tiny purple flowers. You can type all of that into a search box and get back a list of pages about garden pests in general, none of which shows your beetle.

The gap is description. Naming a shape, a shade of green, or a pattern of wing veins takes vocabulary that most of us do not carry around. Experts have that vocabulary, which is why forum answers arrive with terms like pronotum or elytra attached.

A photo skips the vocabulary problem. The image already contains the details you cannot phrase. Lens App reads those details directly, so your job is reduced to holding the camera steady and pressing a button.

How Lens App turns a picture into an answer

You take a photo or pick one from your library. The AI Lens model examines shape, proportion, color, texture, markings, and whatever context the background offers, then returns the most likely identification with a plain language explanation of why it landed there. When two or three candidates are plausible, it lists them instead of pretending to certainty it does not have.

That explanation matters more than the name. "Serrated leaf edges and a square stem" is something you can check yourself in ten seconds. A bare label with no reasoning gives you nothing to verify. This is the difference between a guess you can test and a guess you have to trust blindly.

The same flow handles almost anything you can point a lens at. If you want the mechanics in more detail, the walkthrough on how to [identify an object from a photo](/identify-object-from-photo/) covers cropping, angles, and reading confidence levels, and the broader guide to [identify anything from a photo](/identify-anything-from-a-photo/) shows the range of things people actually submit, from mystery kitchen gadgets to hardware store fasteners.

The categories people ask about most

Living things dominate. A leaf, a flower, a houseplant that is dropping leaves for reasons unknown: the [plant identifier](/plant-identifier/) handles those and usually returns a species plus common care notes. Backyard wildlife and tracks go through the [animal identifier](/animal-identifier/), which is also the tool people reach for with grainy trail camera stills. Anything with six legs and bad timing goes through the [insect identifier](/insect-identifier/), the single most common late-night question we see.

Objects come next. Inherited change, flea market finds, and jars of foreign currency go to the [coin identifier](/coin-identifier/), which reads dates, mint marks, and lettering where they are legible. Beach pebbles, driveway gravel, and that heavy dark chunk someone swears is a meteorite go to the [rock identifier](/rock-identifier/), where luster, grain, and fracture pattern do most of the work.

Outside those buckets, people photograph mushrooms on lawns, mold on bread, spider webs, birds at feeders, engine parts, upholstery fabric, and labels in languages they do not read. All of it starts from the same camera screen.

How to take a photo that gets a specific answer

Fill the frame with the subject. A beetle occupying four percent of a photo of your porch gives the model very little to read. Step closer, or crop before you scan. If your phone struggles at close range, back up slightly and crop after.

Use even daylight when you can. Direct flash flattens texture and blows out the highlights that distinguish one mineral or one wing pattern from another. Shade or an overcast sky is kinder than bright noon sun.

Include something for scale when size is diagnostic. A coin, a fingertip, or a ruler edge tells the model whether it is looking at a two centimeter insect or a ten centimeter one, which can decide the answer entirely.

Shoot more than one angle for anything three dimensional. Leaf undersides, the reverse of a coin, the underside of a mushroom cap, and the profile of a rock all carry information the top view hides. Two decent photos beat one perfect one.

Blurry photos, old photos, and screenshots

Not every question comes from a clean shot. Sometimes the animal moved, the light was bad, or the only record is a photo someone took in 2004 on a phone with a two megapixel sensor. That does not automatically end the attempt. Our notes on whether a [blurry photo is still identifiable](/blurry-photo-still-identifiable/) explain which cues survive motion blur, and the guide to working [from an old, low quality photo](/identify-from-old-low-quality-photo/) covers scans, film prints, and heavily compressed images.

The honest summary: silhouette, proportion, and overall color pattern often survive poor quality. Fine texture, small lettering, and subtle vein structure usually do not. So a blurry photo may get you to a genus or a general family without pinning down the exact species, and that is still useful information.

For screenshots and images pulled from the web, a different question applies. You may want the source rather than the label. A [reverse image search](/reverse-image-search/) traces where a picture appears online, which helps with product listings, art, unattributed photos, and claims that look too tidy to be true.

When a name is not what you actually need

Half the time "what is this" is shorthand for something else. Standing in front of a package of noodles in a store abroad, you do not need the word noodles, you need the ingredient list and the cooking time. Point the camera at the label and [image translation](/image-translation/) reads the text in place, which also works on menus, road signs, appliance instructions, and medication boxes.

Other times the question is really "where do I buy this." You liked a chair in a cafe, a pair of boots on a stranger, a lamp in a hotel lobby. The [shopping finder](/shopping-finder/) matches the item against listings so you get product names and places that carry them rather than a shrug.

Both cases start identically. Take the photo, then choose what you want out of it. Nothing about the capture changes, only the answer you ask for.

Read every answer with healthy skepticism

Lookalikes are real and they are common. Many plants have close relatives that differ in one detail. Wasps and hoverflies borrow each other's stripes. Coins get counterfeited, and worn dates get misread. A confident label from any AI Lens tool, ours included, is a strong lead rather than a verdict.

Two situations deserve extra caution. Never eat a mushroom, berry, leaf, or root because an app named it. Toxic species can resemble edible ones closely enough to fool experienced foragers, and misidentification can cause serious harm. Treat every fungus identification as curiosity only, and consult a qualified local expert if you intend to forage.

Bites and stings are the second case. An identification is not a medical assessment. If someone has a spreading rash, difficulty breathing, severe pain, swelling, or fever after a bite or sting, contact a doctor or emergency services. Do not wait on a photo result, and do not use one to decide against seeking care.

What it costs to use

Lens App is a free download on the App Store with in-app purchases available. Casual use is genuinely free: you get a set of free scans each day on mobile, and one free scan per day on the web version if you would rather work from a browser.

Heavier use is where paid tiers come in. The App Store listing shows several options so people can match spending to habit: weekly, monthly, and yearly subscriptions plus a one time lifetime purchase. Someone identifying a single mystery insect once a month has no reason to pay. Someone cataloguing a coin collection, surveying plants across a property, or translating labels every day during a trip usually does.

Nothing is locked behind a trial that starts before you can see how the tool behaves. Try it on something you already know the answer to. That is the fastest way to judge whether it is accurate enough for the kind of question you plan to ask.

A short workflow that works most of the time

One, get close and fill the frame. Two, shoot in soft light with something for scale if size matters. Three, add a second angle for anything with a top and a bottom. Four, scan and read the reasoning, not just the label. Five, check one stated detail against the object in front of you. If it matches, you probably have your answer. If it does not, crop tighter or reshoot and try again.

That loop takes under a minute and it handles the vast majority of everyday "what is this" moments, from a bug on the windowsill to a stamped mark on the bottom of a mug.

One note on names. Lens App is an independent product, built and operated by our team. It is not affiliated with, endorsed by, or connected to Google or any Google product. People sometimes search for AI Lens or Lens AI App and land here, so it is worth stating plainly.