← The Thread: notes on shopping an outfit from a photo

Comparisons

Which of the six ways finds clothes from a photo?

Six methods, ranked by how often they actually produce a buy link. Some are free, one is slow but accurate, and the right choice depends on what kind of photo you have.

Save to Pinterest

There are six realistic ways to find the clothes in a photo, and they are good at different things. Picking the wrong one is why people conclude that none of this works.

Here they are, with what each one is genuinely strong at and where it falls down.

1. Should you use Google Lens?

Google describes Lens as comparing objects in your picture to other images and ranking those images by similarity and relevance. When one of those matched images is a product listing, Lens can show price and where to buy, drawing on what Google calls a Shopping Graph of more than 45 billion products.

Best at: a clean photo of one current product. A sneaker, a bag, a bottle. If that exact thing is sold in many places, Lens often lands on the listing.

Falls down on: a person wearing five things in a dim room at an angle. You can circle one region to narrow it, but you are doing the splitting by hand.

Cost: free.

2. Should you use Pinterest Lens?

Pinterest says you can search using a photo or part of a photo, then select an object in the results and see similar items for sale on a merchant's site where a match exists. You can also zoom into a specific object inside a Pin to search just that item.

Best at: anything with a strong Pinterest presence. Wedding guest dresses, interiors, styled outfit shots. The taste of the index is genuinely useful.

Falls down on: items nobody pins. You are searching Pinterest, so the answer is bounded by what is on Pinterest.

Cost: free.

3. Should you ask the creator?

Still the only method that can give you a brand name with certainty.

Best at: small accounts. A creator with a few thousand followers often replies, and sometimes replies well.

Falls down on: large accounts. The same question arrives hundreds of times and almost none get answered. You are also waiting, possibly forever.

Cost: free, plus your patience. Ask, then go search the frame anyway rather than waiting on a reply.

4. Should you ask a fashion identification community?

There are communities online whose whole purpose is identifying garments from photos, and the good ones are impressive. Real people who know that a collar shape dates a shirt to a particular decade.

Best at: vintage, obscure brands, heavily patterned pieces, anything where the answer needs knowledge rather than matching. This is where humans still beat software clearly.

Falls down on: speed. You post, then you wait. Post a sharp photo, include a close shot of any tag or hardware, and describe the fabric and the era in your own words. Low-effort posts get ignored.

Cost: free, slow.

5. Should you use the shopping tab in the app?

TikTok and Instagram both have shopping features that can attach products to a post.

Best at: posts where the creator or seller actually attached something. Then it is one tap and you are done, which is unbeatable.

Falls down on: everything else. If nothing was attached, there is nothing to tap, and most posts have nothing attached. It cannot help you with a video of someone who is not selling anything.

Cost: free.

This is what Snagfit does, and it works in the opposite direction from Lens. Snagfit identifies each garment, writes a short text description of it, and searches current shopping listings for that description. One upload returns up to eight separate garments, each with its own results.

Best at: the screenshot case. A person wearing several things, where you want each piece named and priced without circling anything. Also strong when the exact item is gone, because a description still matches similar garments that are in stock.

Falls down on: finding the precise original when many garments look alike. A description of a black leather jacket matches a lot of black leather jackets.

Cost: free to search. Snagfit earns from some retail links, which is written out on the disclosure page.

Why do people conclude that none of it works?

Almost always because they used a method built for one job on a photo that needed a different one, then generalised from that.

Running an image matcher on a blurred frame of five people and getting nothing useful is not evidence that outfit search does not work. It is evidence that image matching needs a clear subject. The reverse is true too: asking a description search to find one precise widely-listed sneaker will sometimes return six sneakers that are almost it.

Match the method to the photo and the hit rate changes completely.

Which two questions decide the method?

Rather than remembering six descriptions, ask two things about your photo.

Is it one item or a whole outfit? One item points at image matching, because you can hand it a clean subject and ask for the same thing back. A whole outfit points at a method that splits the outfit for you, or you will be doing the splitting six times.

Is the item currently on sale somewhere? If yes, most of these methods can find it and you should pick on convenience. If no, image matching has nothing to match against, communities become much more valuable, and a description search becomes the way to find the nearest thing you can actually buy.

Almost every "none of these worked" story is a photo where the answer to the second question was no, and nobody said so.

What can none of the six do?

Worth being blunt, because every one of these methods gets oversold somewhere.

None of them can read a brand that is not in the photo. If there is no visible logo and the garment is not distinctive in shape, the brand is genuinely not present in the image. A tool that returns one anyway has guessed, and a confident wrong brand name is worse than no brand name.

None of them can produce a link to something that is not for sale. Sold-out drops, gifted samples, custom pieces and one-off vintage have no listing. This is the single most common reason a search disappoints, and it has nothing to do with which method you chose.

None of them can fix a bad photo. A blurred frame of a half-hidden garment gives every method the same thin information. Improving the frame helps all six by roughly the same amount.

None of them checks whether the shop is any good. They surface listings. The two minutes of checking before you enter a card number is still yours to do, and it is worth doing.

What does free actually mean here?

All six are free at the point of use. None of them charge you for a search.

That does not mean nobody is paid. Search and shopping tools generally earn from the retail links they surface, Snagfit included: when a merchant link can be wrapped for affiliate tracking, it is, and Snagfit may earn a commission if you buy. It costs you nothing extra and it does not change which results appear or in what order. The full statement is on the disclosure page.

The reason to know this is not suspicion, it is calibration. A tool with a commercial interest in you clicking a listing is still useful, and you should still compare the listings it shows you rather than taking the first one.

Should you combine two methods?

For anything you genuinely want, one method is rarely the best plan.

The pattern that works: run a description search first to get every garment named, then take those names into whichever second method suits the item. Into Lens on a tighter crop if you want the exact listing. Into Pinterest as a text search if you want to see it styled. Into a resale marketplace if the piece is old. Into a community post if it is genuinely obscure, where a garment name plus a sharp photo gets a much better response than a photo alone.

The garment name is the part that transfers between all of them. That is why getting the item named is worth doing even when you intend to buy somewhere else entirely.

So which one do you pick?

One clean product, current season: Google Lens.

Styled outfit shot, interiors, event dressing: Pinterest Lens.

Screenshot of a person wearing a whole outfit: a description-based search, because it splits the outfit for you.

Vintage, obscure, or you need certainty: ask a community and wait.

A piece you care about: do two of them. Run a description search first to get the garment named, then take that name into Lens, a retailer's own search, or a resale app. The words are the part that transfers.

For a closer look at any one matchup, Snagfit vs Google Lens, Snagfit vs Pinterest Lens, Snagfit vs Copped, Snagfit vs Screenshop, and Snagfit vs Amazon StyleSnap go deeper on where each one wins and where it falls down.

Common questions

What is the best way to find clothes from a photo?

It depends on the photo. For a clean shot of one current product, image matching in Google Lens is the most direct route. For a screenshot of a person wearing several things, a tool that names each garment separately, such as Snagfit, gets you further because it breaks the outfit into items you can search individually.

Is there a free way to identify clothing in a picture?

Yes, several. Google Lens and Pinterest Lens are both free to use. Snagfit is free to search as well. None of these charge you for a search, because they earn money from retail links rather than from the person searching, and that arrangement does not change which results appear or in what order.

How does Google Lens find clothes?

Google says Lens compares objects in your picture to other images and ranks those images by similarity and relevance. When a matched image is a product listing, Lens can show shopping details such as price and where to buy. It is image-to-image matching rather than a description search.

How does Pinterest Lens find clothes?

Pinterest says you can search using a photo or part of a photo, then select an object in the results to see similar items for sale on a merchant's site where a match exists. You can also zoom into one object inside a Pin to search that item on its own. Results start from Pins, so the strength of the answer depends on what is pinned.

Does asking the creator work?

Sometimes, and it is slow. A creator with a small audience often replies. One with a large audience gets the same question hundreds of times and answers almost none of them. It costs nothing to ask, so ask, then search the frame yourself instead of waiting.

Are Reddit fashion communities good at identifying clothes?

They can be good, especially on vintage, obscure or heavily patterned items where automated matching struggles. The tradeoff is time. You are waiting on a human, so it suits a piece you care about rather than something you want to check in ten seconds, and a reply can take anywhere from minutes to a day.

Do TikTok and Instagram have their own product search?

Both have shopping features, but they surface items a creator or seller attached to that specific post. If nothing was attached, there is nothing to tap. That is why searching the frame yourself works on far more posts than the built-in tools do, since most posts have nothing attached at all.

What if none of these find the exact item?

Then the item is probably not listed for sale anywhere, which is common with sold-out drops, resale, vintage and gifted product. At that point the useful goal changes from finding the original to finding the closest garment you can actually buy, which is what a description-based search returns.

Sources

Keep reading

Comparisons

Snagfit vs the tagged link: what does a tag cover?

A product tag is the most accurate answer you can get, and it usually covers one garment out of seven. The gap between what is tagged and what is worn is the whole subject.

8 min read
Comparisons

Snagfit vs asking in the comments: which is faster?

Asking a creator where they got something is free and occasionally gets you a direct answer with a link. It also fails silently most of the time, and you find out three days later.

8 min read

← Back to Snagfit