Yandex Reverse Image Search API — Image Matches

Turn one or more public image URLs into structured Yandex reverse-image-search evidence. Each completed lookup produces one compact dataset record containing source-page candidates, visually similar images, Yandex search tags, optional alternate sizes, collection time, and a direct result URL for review.

This Actor is for brand-protection reviews, product and marketplace research, image-provenance research, and content-reuse investigation. Its focused contract is public HTTP(S) image URLs to reverse-image evidence; image uploads, OCR, translation, and identity verification belong in dedicated workflows.

Best fit and workflow guidance

Use this Actor when a source image is the starting point and you need a bounded set of web pages or visual candidates to review. A product researcher can submit a catalog image, inspect sites and similarImages, then send the returned page URLs to a product or seller-enrichment workflow. A rights holder can review where a public image appears, then independently verify context, ownership, licensing, and any action they take.

Choose a text-based image scraper when you already have keywords. Choose an OCR-specific tool when the job is to read visible text. This Actor is deliberately narrow: public image URL in, Yandex reverse-image evidence out.

Quick start

Submit one to ten direct, public image URLs. The default result includes source-page candidates, visually similar images, and Yandex tags; alternate sizes are optional.

{
  "imageUrls": [
    "https://upload.wikimedia.org/wikipedia/commons/e/ed/Elon_Musk_Royal_Society.jpg"
  ],
  "maxResultsPerImage": 20,
  "includeSites": true,
  "includeSimilarImages": true,
  "includeTags": true,
  "includeImageSizes": false,
  "yandexDomain": "yandex.com"
}

Use a direct image file URL rather than a webpage URL. Private-network hosts, local files, IP literals, data: URLs, and login-protected images are rejected before an upstream request. maxResultsPerImage limits each enabled result group independently, from 1 to 50.

Output data

The default dataset has one row per completed source-image lookup. A zero-match lookup is still a meaningful completed record: its result arrays are empty and the terminal outcome is VALID_EMPTY.

Field	Meaning
imageUrl	The public source image submitted for lookup.
searchUrl	Direct Yandex result URL for manual review.
sites	Page candidates where Yandex surfaced the image or a close visual match.
similarImages	Visually similar image candidates with a Yandex result URL.
imageTags	Yandex-provided search suggestions that serve as follow-up research hints.
imageSizes	Optional alternate image URLs and dimensions exposed in the result.
matchedEntity	Optional Yandex result-card context when exposed.
matchCount	Total retained candidates across the enabled groups.
retrievedAt	ISO 8601 collection time.
provenance	Public result-surface and parser version used for the record.

{
  "imageUrl": "https://example.com/catalog/blue-shoe.jpg",
  "searchUrl": "https://yandex.com/images/search?rpt=imageview&url=https%3A%2F%2Fexample.com%2Fcatalog%2Fblue-shoe.jpg",
  "yandexDomain": "yandex.com",
  "matchedEntity": null,
  "sites": [
    {
      "rank": 1,
      "title": "Blue running shoe",
      "url": "https://shop.example/products/blue-shoe",
      "domain": "shop.example",
      "thumbnailUrl": "https://images.example/blue-shoe-thumb.jpg",
      "matchedImageUrl": "https://images.example/blue-shoe.jpg"
    }
  ],
  "similarImages": [
    {
      "rank": 1,
      "imageUrl": "https://images.example/similar-shoe.jpg",
      "title": "Blue running shoe",
      "resultUrl": "https://yandex.com/images/search?..."
    }
  ],
  "imageTags": [{ "label": "blue sneakers", "searchUrl": "https://yandex.com/images/search?text=blue+sneakers" }],
  "imageSizes": [],
  "matchCount": 3,
  "retrievedAt": "ISO-8601 collection timestamp",
  "provenance": { "source": "Yandex Images public result page", "parserVersion": "0.1", "resultSurface": "rpt=imageview" }
}

OUTPUT exposes the stable terminal outcome and billed-event count. RUN_SUMMARY adds safe per-image status, retry, result-coverage, and spend-limit diagnostics. Both records live in the default key-value store.

Pricing and cost control

The configured price is $0.02 for each completed and persisted image search, plus $0.00005 per Actor start. Pay per event (PPE) plus platform usage keeps the search event and compute/proxy usage visible separately; the live Pricing tab is the current source of truth before scheduling production volume.

Set an Apify maximum charge limit to cap a batch. When that limit ends later source-image processing, the Actor retains completed rows and records a PARTIAL outcome with a warning. Input guidance, challenge detection, and result-surface validation keep the dataset focused on completed-search records.

API example

curl -X POST 'https://api.apify.com/v2/acts/khadinakbar~yandex-reverse-image-search-api/runs?token=YOUR_APIFY_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{
    "imageUrls": ["https://upload.wikimedia.org/wikipedia/commons/e/ed/Elon_Musk_Royal_Society.jpg"],
    "maxResultsPerImage": 20,
    "includeSites": true,
    "includeSimilarImages": true,
    "includeTags": true,
    "includeImageSizes": false,
    "yandexDomain": "yandex.com"
  }'

After completion, retrieve the default dataset for the structured records and read OUTPUT for the concise terminal result. The exact Actor path is available after the private deployment step.

AI agent prompt card

Reverse-search this public image URL with Yandex. Keep at most 20 candidates per enabled group, read the dataset after completion, preserve searchUrl and retrievedAt as provenance, treat Yandex tags as research hints, and report the terminal outcome, completed-search event count, and warnings before using a candidate in a decision.

Best results

Use a stable, publicly downloadable image that is at least moderately clear and has a direct URL. Product images, artwork, editorial photos, and catalogue images work best when the subject is visible and not heavily compressed. Start with the default yandex.com surface; test another supplied regional domain only when it reflects the market you are researching.

Returned pages and visual candidates are starting points for rights, source, availability, and context review. Confirm current page content and rights independently when a decision is material.

Responsible use

Use images and resulting pages only when you are authorized to access and collect them, and comply with applicable laws, copyright obligations, and site terms. Keep human review in any high-impact decision involving people. Yandex is a trademark of its respective owner; this independent Actor has no affiliation with Yandex.

Builder's note

I built this Actor around Yandex Images' public rpt=imageview result surface, using direct-image inputs and residential proxy sessions. I added private/local URL rejection before any request, bounded retries, explicit challenge/layout-drift detection, and SDK-coupled dataset writes so a completed-search event always has a complete persisted record. OUTPUT and RUN_SUMMARY make the terminal evidence reviewable without exposing session data.

Related workflows

• Google Images Scraper for keyword-first image discovery.
• Yandex Maps Scraper for location-first Yandex business research.
• Alibaba Listings Scraper when a visual candidate needs marketplace listing details.

Your feedback

If a public image URL produces an unexpected result, share the URL only when you have permission to do so, the selected regional domain, and the terminal RUN_SUMMARY outcome. Keep API tokens, cookies, and private image URLs out of support requests.