Google Lens Reverse Image Search — Visual Matches, No Login

Paste links to pictures and get Google Lens's answer for each one as flat rows: every visual match (the page, its title, the site, the original picture and its size), the pages about the picture and its short videos — or every page carrying that exact picture. Priced per image searched, from $4.50 per 1,000 image searches. No login, no API key.

Marketers, brand-protection and e-commerce teams use reverse image search to find where a picture is used, who else sells a product photo, which sites copied an image without attribution, and where a picture first came from. This actor runs Google Lens on your image links and returns one row per result, so a list of product photos or campaign images becomes one table you can filter, sort and export.

Using an AI agent? Pin this actor in Apify's MCP server with one link: https://mcp.apify.com?tools=steadyfetch/google-lens-reverse-image-search

• Actor id: steadyfetch/google-lens-reverse-image-search
• Input: { "imageUrls": ["https://example.com/product-photo.jpg"], "searchType": "visual" }
• Cap the bill: set maxTotalChargeUsd on the run (a run option, not Actor input), e.g. 0.50 — the run searches only the images that can pay.
• Price: $0.0045 per image searched at Gold ($0.006 on the Apify free plan), no start fee. Full table below.
• How often it changes: Google Lens finds new pages carrying a picture every day. A weekly re-search of the same image, with results you already received marked seenBefore, shows what is new.

Unofficial. Not affiliated with, endorsed by, or sponsored by Google.

This actor finds where an image appears on the web. It is not a people-search tool: do not use it to identify a person from a photo.

What you get

One flat row per result (rowType: "result"), in the order Google Lens shows them:

Column	What it holds
imageUrl · imageIndex	the image link you gave, and its place in your list
resultType	visual_match (a picture like yours), page (a page about the picture), video (a short video), or exact_match (a page carrying your exact picture, exact mode)
position	its rank within its type
title · pageUrl	the page's title and link
sourceName · sourceDomain	the site's name and domain
originalImageUrl · originalImageWidth · originalImageHeight	the matching picture's own address and its size in pixels (visual matches); in exact mode, the size the picture is shown at on that page
thumbnailUrl	Google's small preview of the match
description · price · currency	the page's snippet, and a price where the page is a shop listing
seenBefore	true when this actor already delivered this result to your account for the same image link

Then one image row per searched image (rowType: "image") with its counts — results found and delivered, visual matches, pages, videos, exact matches — and the image's one search fee. An image Google Lens found nothing for gets its image row too, saying "Google Lens found no matches for this image."

Every row also carries charged, chargedEvents, status and statusReason, so you can reconcile your bill from the dataset itself.

How to use it

1. Put your picture links in Image links — one per line, the address of the picture itself (https://…/photo.jpg), not of the page it sits on. Up to 100 per run.
2. Pick the Search type: visual (pictures like this one, about 60 a picture, plus pages and videos) or exact (every page carrying this exact picture, up to about 400). Google Lens's own words are read too, and the first row says how: search, similar, products, shopping, all, image, reverse image or lens run visual; exact match, same image, identical, find image source or source run exact. Lens's text, translate, about this image and homework modes are not image searches, so they are refused on each image's row and nothing is charged.
3. Press Start — or schedule it weekly.

Optional: Max results per image (the fee is the same whatever it is set to), Only results I have not received, Max run seconds.

Refused before any search, and never charged: an entry that is not a public link to a picture, a pasted file (data: link), a link that opens a web page, a link with nothing at the address, and a picture larger than 7 MB. Each gets its own row saying why.

Pictures Google Lens cannot open

Google Lens fetches each picture from its own address. Some sites refuse it, mostly for large pictures. When Lens answers nothing for a picture this actor can open, the run copies the picture into your run's own storage (in your Apify account), sends Lens a private signed link to that copy, and deletes the copy once Lens has answered. The image row and its results say when this happened (route: "copy"; otherwise route: "direct"). The second search is on us: one image, one fee. The Standby endpoint makes no copy; it answers an uncharged note and a run re-sends the picture.

Only the new results

This actor keeps a small record per image link in a store named google-lens-account in your Apify account: the ids of the results it delivered to you, nothing else. Results you already received are marked seenBefore: true; turn on Only results I have not received to leave them out. Every search is still a fresh Google Lens search and is charged as one.

Pricing

Pay per image searched, all-inclusive (no platform usage on top), no start fee:

Event	Apify Free plan	Bronze	Silver	Gold and up
Image search (one image Google Lens answered, matches or none)	$0.006	$0.0055	$0.005	$0.0045

An image with no match is still charged, because it was searched. Never charged: an image refused before the search, one Lens did not answer in time, one it returned nothing for that could not be re-sent, and the bare-Start sample.

Examples at Gold: 1 image = $0.0045. 10 images = $0.045. 100 images in exact mode = $0.45, however many pages each one returns.

Good to know

• A time limit ends the searching, never the delivering: images already searched are delivered, and the receipt says which limit stopped the run.
• Free plan: on the Apify free plan a run searches five images per 24 hours; the row says when it resets. Paid plans have no daily allowance.
• Standby endpoint: one image per HTTP request (GET /?imageUrl=https://example.com/photo.jpg), at the same price per image. It keeps no account memory and makes no copy of a picture.
• Results drift: Google Lens is read as it answers in the United States in English, and on a famous picture it still mixes pages in other languages.
• Bare Start: with no image set, the run searches one public-domain painting as a sample and charges nothing.
• Honest rows: not_found (nothing at the address), no_data (Lens returned nothing and the picture could not be re-sent), try_again (no answer in time — a re-run fixes it). None is charged.
• The receipt is the last row (run_summary) and the run's OUTPUT record.

Support

Questions, an image that returns something odd, or a column you need: open an issue on the Issues tab and it will be looked at.

If it earned its keep, a rating helps other buyers find it, and saving the actor keeps it one click away.