How to Find Where an Image Appears Online With One API

When someone reposts your product photo or lifts your artwork, the first question is always the same: where else is this image online? Answering it by hand means uploading the picture to a reverse image engine, scrolling the results, and copying links into a sheet, one image at a time. I wanted a version I could call like an API and run across a whole folder of images at once, so I built one.

What the Yandex Reverse Image Search API does

The Yandex Reverse Image Search API takes a public image URL and returns every page where that image appears, plus visually similar images, other resolutions, product matches, and descriptive tags, all as structured JSON. It queries Yandex's reverse image engine, which is widely regarded as the strongest on the open web for finding re-uses and earlier versions of a picture.

What it returns

Every row carries a result_type label, and toggles control which kinds a run returns:

• matching_page: page title, link, and snippet for every page hosting the image.

• similar_image: thumbnails and source links for visually similar content.

• image_size: other resolutions of the same file, large to small.

• shopping_result: e-commerce listings that match, with prices where available.

• image_tag and knowledge_graph: descriptive terms and an entity card when the subject is recognizable.

Find matching pages in three steps

Open the Actor, paste a public image_url, keep the matching pages toggle on, and run it. Download the dataset as JSON, CSV, or Excel. Billing is pay per result with a cap you set, so a 20-row test run costs very little. The task Find where an image appears online with Yandex runs it in one click.

Call it as a REST API

Kick off a run with a single POST to the Actor's run endpoint:

POST /v2/acts/johnvc~yandex-reverse-image-search/runs
{ "image_url": "https://example.com/product.png", "max_results": 20 }

The run writes one JSON row per result to its dataset, which you then read back over the API. The task Reverse image search with Yandex, results as JSON is this call saved and runnable.

Wire it into Claude with MCP

The Actor runs over MCP, so Claude, Claude Code, and Cursor can call it as a tool. Paste an image URL into a conversation and ask where else it appears online, and the agent returns structured rows with no scraping code. The task Reverse image search in Claude via MCP has the setup.

What people build with it

Photographers and brands track unauthorized use of their images, fact-checkers trace a photo to its first appearance, and e-commerce teams hunt marketplace listings that copy their product photos. Filtering on result_type turns each section into its own dataset. See all the runnable examples on the Peerlist project page.

FAQ About Reverse Image Scraping

Is there a free reverse image scraper?

This one bills per result row, and max_results caps a run before it starts, so a small test run costs very little. New Apify accounts include free platform credit that covers your first investigations.

How do I call the reverse image API from code?

Send a POST to the Actor's run endpoint with an image_url and an optional max_results, then read the dataset back as JSON, CSV, or Excel. Each row is keyed by result_type, so you filter to matching pages, similar images, or shopping matches in a line of code.

Can I run reverse image search from Claude with MCP?

Yes. The Actor is exposed as an MCP tool, so an AI agent can call it directly. Paste an image URL and ask where it appears online, and the model gets structured rows back from Yandex, no code required.

What are the limits of this reverse image scraper?

The image must be publicly reachable over http(s), since Yandex fetches it; localhost and private links will not work. Results reflect Yandex's index and can differ across its six regional domains, and its face-matching strength deserves the obvious caveat: use it responsibly and lawfully.

Questions or a coverage gap? Open an issue on the Actor page and I usually reply within a day.