Yandex Images Scraper

Search Yandex Images and export normalized, source-backed image records from the public structured metadata embedded in result pages. The Actor maps only documented output fields and never stores opaque request identifiers, signatures, compressed state, cookies, or raw responses.

Input

• mode: single for one query or batch for multiple queries.
• query and queries: primary and batch search terms; duplicates and empty values are removed.
• maxItems, maxPages, and maxConcurrency: global record limit, per-query page limit, and parallel query controls.
• language, country, and region: output locale, optional country, and Yandex numeric lr region.
• familyFilter, size, color, imageType, and orientation: website-native search filters.
• maxRequestRetries, navigationTimeoutSecs, requestHandlerTimeoutSecs, and resultsWaitSecs: bounded reliability controls.
• proxyConfiguration: optional Apify Proxy or custom proxy settings. Direct access is the default.

{
  "mode": "batch",
  "queries": ["кот", "собака"],
  "maxItems": 20,
  "maxPages": 3,
  "maxConcurrency": 2,
  "language": "ru",
  "country": "RU",
  "region": "213",
  "proxyConfiguration": { "useApifyProxy": false }
}

Output

The 51-field contract includes stable IDs, Yandex image/document IDs, full and thumbnail image URLs, real publisher page URLs and domains, titles and descriptions, full and thumbnail dimensions, pixel count, file size, format, MIME type, animation/censorship flags, freshness classification, pagination/query position, locale, search URL, and ISO-8601 scrape time.

Required fields are validated before records are pushed. Malformed metadata, missing publisher pages, duplicates, challenge pages, invalid URLs, and empty values are excluded. Explicit no-results pages complete successfully with an empty dataset; selector drift fails clearly.

Extraction and resilience

The primary extractor parses Yandex's ImagesApp structured state and maps selected public fields. A DOM fallback is retained for layout changes. Crawlee sessions and consistent Chromium fingerprints are enabled. Temporary network, proxy, 5xx, 403, and 429 failures use bounded retries; explicit CAPTCHA pages fail closed and are not retried or stored.

Verification

npm ci
npm test
npm run lint
npm run validate:dataset
apify validate-schema
apify run --purge --input-file INPUT.json

Representative single, batch, filtered, and no-results inputs are in qa-inputs/.