Reverse Image Search API

What does Reverse Image Search API - Google Lens do?

This actor is a reverse image search API that takes any image URL, runs it through Google Lens, and returns every exact-match page Lens found, including the page URL, title, publisher, publication date, and thumbnail. Google Lens no longer returns the direct URL of each matched image — only the page where it was found. To bridge that gap, this actor can optionally visit each result page and find the original-source image URL, so you get a direct download link instead of just the page address.

It is a clean, JSON-first Google Lens API for developers, OSINT teams, brand-protection workers, photographers, and e-commerce operators. Try it from the Apify Console, call it from the Apify API, schedule it as a cron job, or wire it into n8n, Zapier, Make, or your own pipeline. Apify handles residential proxy rotation, retries on 429 / /sorry/ blocks, scaling, and monitoring.

Unlike most reverse image search tools, this one searches the whole frame instead of an auto-cropped subsection of your image. The result is true exact matches, not visually-similar suggestions.

Why use this Google Lens reverse image search API?

• Reverse image search at scale. Bulk / batch reverse image search across hundreds or thousands of source images per run. Output streams to the dataset as it lands. Export to JSON, CSV, Excel, or HTML.
• Finds the original image source. A Google Lens exact-match result set contains only Google-hosted thumbnails — in a typical run, several hundred encrypted-tbn thumbnails and zero links to the actual image files. Other reverse image search tools hand you that page URL and stop. This actor goes to each page and finds the real image file, so you get a downloadable original instead of a thumbnail.
• Exact-match focus. Full-frame Lens queries cut out the noise of visually-similar matches, so what you get back is genuinely the same image. Exactly what you need for image-source lookup, copyright takedowns, and image-theft detection.
• Pay only for what you get. Pay-per-event pricing: one tiny charge per Lens search, one per match returned. No subscription, no monthly minimum, no idle compute units — run a single image to try it and pay for that one image. Unresolved matches aren't charged.
• No CAPTCHA dance for you. Apify residential proxies rotate per attempt, with automatic retry on a fresh exit IP for any blocked request.

How to find the original source image URL

Google Lens tells you where an image appears. It does not tell you which file it found. Every match comes back with a Google-hosted thumbnail and a page address — nothing that points at the image itself.

That is the gap this actor closes. With resolve enabled, each result page is opened, every candidate image on it is examined, and the one that is genuinely your image is returned as resolved_image_url. You get a direct, downloadable link to the full-size original on the publisher's own server.

This is what makes the output usable as evidence rather than as a lead. A page URL tells you someone used your photo; the original file is what you attach to a DMCA notice, hash into a database, compare against your master, or archive before it disappears.

How each image source match is verified

Finding an image on a page is easy. Proving it is your image is the hard part, and it is what separates a real image source finder from a scraper that grabs the largest file it can see.

Every candidate is compared against the Google Lens thumbnail using perceptual hashing — a fingerprint of the visual content that survives resizing, recompression, and format changes. Only images that actually match are returned, and each one carries its confidence level in resolve_status:

Status	Meaning
match_definitely	High-confidence match — the file is your image
match_maybe	Likely match, slightly outside the strict threshold
match_no	Page loaded, but the original image was not found on it
no_candidate	Page loaded but had no usable images
fetch_failed	Paywall, captcha, or network block on the destination page
thumbnail_failed	The Lens thumbnail could not be loaded, so the page was not tried — the page URL is still kept

Because the check is visual rather than filename- or size-based, it works on sites that rename, crop, watermark, or re-encode the images they republish. You never have to manually confirm that a returned file is the right one.

How to use this reverse image search Actor

1. Open the actor in the Apify Console.
2. Paste one or more image URLs into the Image URLs field. Direct links to .jpg, .png, .webp, .heic, or .avif files all work, and bigger originals give cleaner exact matches.
3. Leave Resolve original-source image URL turned on (recommended) for the full pipeline, or turn it off for a faster, cheaper run when you only need the page URLs and titles.
4. Click Start.
5. Watch results stream into the Dataset tab. Download as JSON, CSV, Excel, or HTML when the run finishes, or fetch them via the API while the run is still going.

To automate this, hit the Apify API directly or use one of the official client libraries (Python, JavaScript). Every run accepts the same JSON body as the Input tab.

Reverse image search API example — Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")

run = client.actor("thodor/google-lens-exact-matches").call(run_input={
    "imageUrls": ["https://example.com/photo.jpg"],
    "resolve": True,
    "country": "US",
})

for match in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(match["page_url"], "->", match.get("resolved_image_url"))

Reverse image search API example — JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });

const run = await client.actor('thodor/google-lens-exact-matches').call({
    imageUrls: ['https://example.com/photo.jpg'],
    resolve: true,
    country: 'US',
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((m) => console.log(m.page_url, '->', m.resolved_image_url));

Both return the same rows as the Dataset tab, so you can prototype in the Console and move to code without changing anything.

Input

Field	Type	Required	Description
imageUrls	array of URLs	yes	Image URLs to reverse-search. Accepts bare strings or {"url": "..."} entries.
resolve	boolean	no (default true)	When true, each match is enriched with the original-source image URL — when we can still find it on the destination page. Sometimes the image has been moved or deleted, and you only pay for the matches we successfully resolve.
country	string (ISO 3166-1 alpha-2)	no (default US)	Country used both as the Apify residential proxy exit region for every Google Lens request and as the gl= parameter sent to Lens (biases the regional index).

Example input JSON:

{
  "imageUrls": [
    "https://example.com/photo.jpg",
    "https://example.com/another.png"
  ],
  "resolve": true,
  "country": "US"
}

Output

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel, or read it as a spreadsheet directly from the Apify Console.

One dataset record per Lens match. A single input image typically produces several to several dozen output rows:

{
  "source_image_url": "https://example.com/photo.jpg",
  "page_url": "https://news.example.org/article/123",
  "title": "Original story about the photo",
  "publisher": "Example News",
  "published_date": "2024-09-12",
  "lens_thumbnail_url": "https://encrypted-tbn0.gstatic.com/...",
  "lens_expected_width": 1920,
  "lens_expected_height": 1080,
  "scraped_date": "2026-05-14",
  "resolved_image_url": "https://news.example.org/wp-content/uploads/2024/09/photo-full.jpg",
  "resolve_status": "match_definitely",
  "page_status": 200,
  "thumbnail_mismatch": false
}

Data fields

Field	Description
source_image_url	The image you searched
page_url	Web page where Google Lens found this match
title	Page title as Lens displays it
publisher	Publisher / site name
published_date	Publication date when Lens surfaces one (ISO 8601)
lens_thumbnail_url	The thumbnail Lens shows for this match (usable as-is)
lens_expected_width / lens_expected_height	Dimensions Lens reports for the match
scraped_date	Date this actor row was produced
resolved_image_url	The original-source image file on the destination page (present when resolve=true)
resolve_status	match_definitely, match_maybe, match_no, no_candidate, fetch_failed, or thumbnail_failed
page_status	HTTP status of the destination page fetch
thumbnail_mismatch	true when the resolved image didn't match the Google Lens thumbnail for this row, but matched a different Google Lens thumbnail from the same search

Use cases for reverse image search

OSINT reverse image search and investigation. Trace where a photograph first appeared online, who has republished it, and on what dates. Investigators get the full-resolution original image source, not just the Lens thumbnail. That is critical when verifying authenticity or building a timeline.

Dating profile verification. Run a list of profile photos through the actor to catch stolen or recycled images. Reverse image search for dating profiles works across Tinder, Hinge, and every other major dating app, anywhere Google has indexed the photos. Useful for both individual safety checks and Trust & Safety teams running bulk reverse image search for dating sites.

Scammer and catfish detection. A scammer's profile picture often originates from a stock photo, a public Instagram account, or another victim's social media. Reverse image search scammer photos to surface those origins and prior fraud reports.

Copyright protection, artwork, and image-theft detection. Photographers, illustrators, and stock-photo licensees use bulk reverse image search artwork to find unauthorized reuse. Used as an image copyright checker, it finds every page carrying your work and returns the file each one is serving — downloadable evidence URLs straight from the dataset, ready for your DMCA workflow. Finding the original image source on every infringing page is exactly what this actor is designed for.

E-commerce product matching. Reverse image search clothing, an outfit, furniture, jewelry, or any product photo to find duplicate listings, knock-offs, and competitor pricing across marketplaces.

Social media monitoring and brand protection. Track image usage online across every major platform, and find where a brand asset, leaked screenshot, or viral photo has spread.

Specialized lookups. Reverse image search plants, mugshots, art, anime stills, or anything else where you want the authoritative page of context, not just visually-similar images.

Supported platforms and sources

Google Lens indexes the public web, so any page reachable by Googlebot can surface as a match. In practice you'll routinely see results from:

• Social media: Facebook, Pinterest, Twitter / X, Instagram, TikTok, YouTube, LinkedIn, Tumblr, Reddit, VK
• Marketplaces: eBay, AliExpress, Taobao, Depop, Shutterstock, Etsy
• Image and art hosts: DeviantArt, Pixiv, Flickr, Imgur, ArtStation
• News and publisher sites: every major outlet plus long-tail blogs
• Dating sites and forums where Google has indexed the photo

The actor does not separately query each platform's own search endpoint. It leverages Google Lens' single, consistent index of all of them. That means a single reverse image search call effectively gives you a Facebook reverse image search, Pinterest reverse image search, Twitter reverse image search, TikTok reverse image search, Depop reverse image search, eBay reverse image search, and so on, wherever Google has crawled.

Note: private content (private Instagram, closed Facebook groups, WhatsApp messages, Discord servers) is not indexable by Google and therefore not reachable by Lens or this actor.

Tips and advanced options

• Disable resolve for speed and cost savings. With resolve off, each input image finishes in a few seconds. With resolve on it can take 30-90 seconds per image because the actor visits every result page to find the original image. Turn it off when you only need page URLs and titles.
• Use big, clean source images. Lens is most precise with originals ≥1000px on the long edge. Tiny thumbnails or heavily-cropped versions produce more visually-similar noise and fewer true exact matches.
• Interpret resolve_status correctly. See How each image source match is verified for what each value means.
• thumbnail_mismatch=true means the resolved image didn't match the Google Lens thumbnail for this row, but did match a different Google Lens thumbnail from the same search. Usually happens when Lens displays a cropped or scaled thumbnail and the page hosts a clean original — the match is still real.
• Schedule it. Use Apify Schedules to monitor image leaks or stolen art daily / weekly. Pair with Apify Webhooks to notify Slack, Discord, or email whenever new matches show up.
• Combine with other actors. Pipe the output into a downstream actor that screenshots page_url, runs OCR, or scores publisher against a reputation list.

FAQ

Does this resolve images from the following domains? Yes, across the public web. From real runs we routinely get resolved image URLs from:

• Social: Facebook, Instagram, Threads, TikTok, X / Twitter, Pinterest, LinkedIn, YouTube
• Encyclopedic: Wikipedia (all language editions), Wikimedia Commons, Britannica
• News & media: BBC, Forbes, Reuters, CNN, CNBC, India Today, Hindustan Times, The Daily Beast, and most other major outlets
• E-commerce: Amazon, eBay, AliExpress, Shutterstock
• Long tail: Medium, Substack, WordPress blogs, fandom wikis, personal sites

What doesn't resolve cleanly:

• Private or login-walled content — private Instagram accounts, closed Facebook groups, Discord, WhatsApp, paywalled articles. If Google can't index it, we can't find it
• Sites behind aggressive bot protection — parts of Reddit, namu.wiki, and certain regional outlets sometimes return fetch_failed

Does Google Lens give you the original image URL? No. Google Lens returns the page where an image was found and a Google-hosted thumbnail (encrypted-tbn0.gstatic.com), never a link to the image file on the publisher's server. A single exact-match result set typically contains several hundred of those thumbnails and no source image URLs at all. To get the actual file you have to open each result page and identify the right image on it, which is what this actor's resolve option does automatically.

How do I find the original source of an image, not just the page? Run the image through this actor with resolve enabled. Each match returns resolved_image_url — a direct link to the full-size file on the site that published it — alongside the page URL, title, publisher, and date. Every returned file is perceptually matched against the Lens thumbnail first, so you are not guessing which image on the page was the one Lens saw.

Why don't I get results for some images? Either Lens has no exact match in its index, or every proxy attempt was blocked. The actor logs both cases distinctly. Try a different version of the source image (typically a larger or less cropped one).

Why does the resolved image sometimes differ in size from the Lens thumbnail? Publishers often serve compressed, resized, or watermarked thumbnails on listing pages. The resolver looks for the largest, highest-priority visually-matching image on the page. That's typically the original asset, not the same crop Lens showed.

Is there a free way to try this reverse image search API? Yes — Apify's free plan includes $5 of platform credit every month, and a trial run costs a tiny fraction of that. There's no subscription and no monthly minimum, so you can run a single image, see the output, and stop. You're never billed for idle compute, and matches that fail to resolve aren't charged at all. See the Pricing tab for current rates.

Is reverse image search legal? Searching an image against a public index is the same activity as using Google Lens or TinEye in a browser, and this actor only reaches pages Google has already crawled and indexed — nothing private, logged-in, or paywalled. What matters legally is what you do with the results. Common uses like copyright enforcement, DMCA evidence gathering, brand protection, and catfish or fraud checks are exactly what reverse image search exists for. If you process images of identifiable people, handle the output under whatever privacy rules apply to you (GDPR, CCPA), and follow each destination site's terms before republishing anything you find. This actor returns links and metadata — it does not host, alter, or redistribute anyone's images.

Can I use this for a video or a GIF? Lens treats animated GIFs as their first frame. Reverse image search on a video isn't directly supported. Extract a representative frame yourself and feed it as a still image.

How does this compare with other Apify Google Lens actors? This actor specializes in exact-match reverse image search with original-image URL resolution. Other Google Lens actors on Apify focus on visually-similar matches, OCR, translation, or AI-mode Lens features. If you need OCR or translation, look at those. If you need the page URL and the actual downloadable original image for each exact match, at scale, with structured JSON output, this is the right tool.