A photo is everywhere this morning: a flooded city, a celebrity in an unlikely place, a politician caught in a strange moment. Before you share it, you want to know two things. Has this picture existed before, somewhere else, with a different story attached? And was it captured by a camera at all, or generated by SI (AI)? A reverse image search answers the first question better than any other free tool. It only partly answers the second.
This guide shows how to run one on a computer and on a phone, which engine suits which job, how to find the first appearance and the original version of a picture, and why the method struggles with brand new SI images. Button names were checked on each service's own help pages on the day of publication; interfaces change often, so look for the closest equivalent if a label differs.
What a reverse image search actually does
A normal search starts from words. A reverse image search starts from a picture and asks the engine to find other places where that picture, or something close to it, appears online. TinEye describes the core of the technique well: it creates a compact digital "fingerprint" of your image and compares it with every image in its index. It does not read the file name or the metadata, and it can match copies that have been "cropped, edited or resized".
Google Lens goes further. Its results can include web pages that contain the image or a similar one, similar images, results about objects recognised in the picture and, in some regions, an SI overview. That breadth is useful for identifying a plant or a landmark, but it also means "similar" results can look like matches when they are not. Keep that distinction in mind: an exact or edited copy of your picture is evidence; a picture that merely looks alike is not.
How to reverse image search on a computer
Google: upload, drag, paste a link or right-click
According to Google's help page, visual search with Lens works in most browsers, including Chrome, Firefox, Safari and Edge. You have four ways to start:
- Upload a file. Go to google.com, click Search by image, then Upload a file, select the image and click Open or Choose.
- Drag and drop. Find the image file on your computer and drag it into the Google search box.
- Paste a link. Right-click the image on the website, choose Copy Image Address, go to google.com, click Search by image, paste the address into "Paste image link" and click Search.
- Right-click in Chrome. On any website, right-click the image and choose Search with Google Lens. The results open in a sidebar.
After uploading, you can type words into the "Add to your search" box to narrow the results.
Google's "About this image"
For pictures that already appear in Google Images, there is a dedicated context panel. Go to images.google.com, search, select the image, then click More and About this image. When Google launched the feature in October 2023, it said the panel shows when an image or similar images may first have been seen by Google Search, how other pages, including news and fact-checking sites, use and describe it, and any metadata the creator added, including fields that may say it was generated or enhanced by SI. Google's help page on image details adds that this information can come from IPTC fields, from C2PA Content Credentials and from SynthID, Google's invisible watermark. If you want to understand those two signals, read our guide to C2PA and Content Credentials.
Bing Visual Search
Microsoft's Visual Search page lets you drag and drop an image, take a photo, paste an image or a URL, or browse for a file. In the Edge browser on desktop, you can right-click and select the Visual Search icon, then frame the part of the page you want to search. On Windows, the Photos app has a "Visual search with Bing" option, and Microsoft offers Bing Visual Search extensions for Chrome and Firefox.
TinEye
TinEye accepts an uploaded image, a pasted image or an image URL, and you can drag and drop a file onto the search page. If you use Firefox, Chrome, Edge or Opera, its official extension adds a right-click option, Search Image on TinEye. In the extension's options you can choose the order of results: Best Match, Most Changed, Biggest Image, Newest or Oldest. TinEye says it never adds your uploaded image to its index and keeps it only briefly to show your results.
How to reverse image search on a phone
Android
Google's instructions for Android use the Google app or the Chrome app:
- An image on a website: open the page in the Google app or Chrome, touch and hold the image, and tap Search image with Google Lens. In Chrome, this requires Google to be your default search engine.
- An image already on your phone (a screenshot or a picture received in a chat): open the Google app or Chrome, tap the Google Lens icon in the search bar, and select the photo under "Screenshots".
- Part of an image: drag the corners of the selection box around the area you care about.
iPhone and iPad
On iOS, Google lists the Google app, the Chrome app and Safari:
- An image on a website: in the Google app or Chrome, touch and hold the image. In the Google app, tap Search inside image; in Chrome, tap Search image with Google. Chrome again needs Google as the default search engine.
- An image in your camera roll: tap the Google Lens icon in the search bar, then tap Photo picker and select the photo. You can then drag the corners of the box to search only part of it.
TinEye has no mobile app, but its website works on phones: open it in your browser and upload a file or paste an image address, exactly as on a computer.
Screenshot or original? A screenshot from a social app adds borders, captions and interface elements, and often lowers the resolution. Crop to the picture itself before searching, or better, save the image from the post and search that file.
Which engine for which job
| Engine | Best at | Watch out for |
|---|---|---|
| Google Lens / Google Images | Huge index, objects and places, pages that use the image, "About this image" context and metadata | Many "similar" results that are not the same picture |
| Bing Visual Search | A second large index, convenient on Windows and in Edge | Results lean towards similar images and shopping |
| TinEye | Exact and edited copies, sorting by Oldest, results kept after pages disappear | Smaller index, social media rarely indexed, no "similar image" search |
| Yandex | Often strong on faces and landscapes, according to investigators | That ranking comes from a 2019 comparison; confirm what it finds elsewhere |
| InVID-WeVerify plugin | Sends one image to several engines at once; extracts keyframes from videos | Browser extension aimed at professionals, more options than a casual user needs |
A few points behind that table. TinEye itself says Google's index is larger than its own, but argues that its matches have fewer false positives, that it finds cropped and modified copies, and that it "doesn't delete old results" and shows when they were found. It also says it does not usually index Facebook, Instagram or X images, because most social networks prohibit it, and that it cannot recognise the contents of a picture: it will not find a different photo of the same person. The open-source investigators at Bellingcat compared the engines in a 2019 guide and found Yandex the strongest on faces and landscapes, Google best for well-known people and widely shared pictures, and TinEye focused on exact duplicates. Engines have changed since, so treat that as a starting point.
The InVID-WeVerify verification plugin, designed and maintained by AFP's Medialab, bundles this into one tool used by journalists and fact-checkers. Its documentation describes reverse image searches on Google, Bing, Yandex, TinEye and Baidu, plus the ability to split a video into keyframes that can then be searched like still images. That matters for viral clips, as our SI video detector page also notes: a frame from a "new" video often turns up in an older one.
How to verify a viral image, step by step
- Get the best copy you can. Save the image rather than screenshotting it, and crop out captions or watermarks added by reposting accounts.
- Search at least two engines. Google Lens or Bing for breadth, TinEye for exact copies. If nothing comes up, search a cropped part: a sign, a building, a distinctive object. Bellingcat suggests cropping manually and keeping the crop large enough (it mentions more than 200 by 200 pixels) when the detail is small.
- Look for the first appearance. In TinEye, sort by Oldest. In Google, open "About this image" to see when it was first seen. A picture "from yesterday's storm" that was online three years ago is settled.
- Look for the original version. Compare sizes and framing. The original is usually larger, less compressed and wider. TinEye's Biggest Image and Most Changed sorts help here: a wider frame can reveal what a crop was hiding, and a "most changed" copy can show where text or objects were added.
- Read the pages, not just the thumbnails. Who published it first? A news agency, a photographer's portfolio, a stock library, an SI art community? Snopes recommends checking press and stock libraries such as Getty Images, reliable outlets, and existing fact checks.
- Check the account that posted it. In one case, Snopes traced an alleged photo of Pope Leo XIV with Bad Bunny to its original Instagram post, whose caption carried an "#AI" hashtag. The source said it plainly; the reposts dropped that context.
What a reverse image search cannot tell you
The method has one blind spot that matters more every month: a brand new SI image has no history. If someone generated a picture an hour ago and posted it first, there is nothing earlier to find. The engines will happily return "visually similar" pictures, which proves nothing either way. An empty result does not mean an image is real; it only means the engines have not seen it before.
Other limits are worth knowing:
- Indexes are incomplete. TinEye says plainly that if it misses an image, "it is simply because we have not yet crawled" the page. Private groups, messaging apps and many social networks are partly or entirely invisible to search engines.
- Edited real photos are a mixed case. If someone used SI to change part of a real photo, a reverse search can find the untouched original, which is the strongest possible evidence. If the edit is new and the original was never published, you are back to an image with no history.
- Old does not mean authentic. SI images have circulated for years. First appearance tells you about context, not about how a picture was made.
- Metadata is often missing. "About this image" can only show C2PA or IPTC information if it survived. Most social platforms re-encode uploads, and screenshots carry none of it.
Combine it with an SI detector
This is where the two approaches complement each other. A reverse image search looks outward, at where a picture has been. An SI detector looks inward, at the file itself: signed Content Credentials, IPTC "digital source type" tags, camera EXIF data and the statistical traces that generators leave in pixels. Our SI image detector reports those clues separately from its score, so you can see why it leans one way.
Use them together. When a reverse search finds an older original, you mostly have your answer about context, whatever a detector says. When it finds nothing, the detector's clues become more useful, especially metadata, which is close to proof when present. And when both are inconclusive, say so. Detectors make mistakes in both directions, as we explain in are SI detectors accurate?, and no single score should decide whether you publish, share or accuse.
The same combination protects you outside the news. A profile picture found on a stock site or another person's account is a classic warning sign of fraud; our guide to SI scams with cloned voices, fake photos and fake profiles covers the other checks. For the visual clues you can look for yourself, such as hands, text, reflections and light, see how to spot SI images, and if you want to practise, our SI or Not? game puts your eye to the test.
The short version
- On a computer: google.com, then Search by image; or right-click an image in Chrome and choose Search with Google Lens. Repeat on TinEye and Bing.
- On a phone: touch and hold the image in the Google app or Chrome, or tap the Lens icon and pick the photo from your device.
- Look for the oldest copy and the largest, least cropped version, then read who published it first.
- No results is not proof of anything. Fresh SI images have no past, which is why a reverse search and an SI detector work best side by side.
