Bing Visual Search Scraper

Submit a public image URL to Bing Visual Search and collect the subject Bing identifies, source pages, and image matches exposed on the result page.

Input

Input	Purpose
imageUrl	Direct public HTTP(S) image URL
market	Bing market such as en-US or en-GB
maxResults	Maximum unique records to store
maxPages	Maximum Bing result pages to inspect
maxRequestRetries	Retries for temporary browser and navigation failures
navigationTimeoutSecs	Browser navigation timeout
requestHandlerTimeoutSecs	Maximum time for image upload and result extraction
imageDownloadTimeoutSecs	Input image download timeout
maxImageBytes	Maximum downloaded input image size
headless	Run Chrome without a visible window
proxyConfiguration	Optional authorized Apify Proxy or custom proxy settings

The Actor downloads and validates the image, opens Bing's public Visual Search page for the selected market, and submits the image bytes through Bing's upload control. The input URL's query string is not sent to Bing; searchImageUrl is stored without query parameters. It uses the page's identified search query to extract result cards and follows Bing outbound redirects to retain direct source-page URLs.

Output

Records use three resultType values:

• identifiedObject: Bing's detected subject and its overview description when Bing exposes one.
• sourcePage: a source page returned in Bing's web results, including title, snippet, domain, date text, and any image URLs exposed on that card.
• visualMatch: an image match exposed in Bing's visual-search module, including the original image URL, thumbnail, dimensions, Bing image ID, and source page when available.

Every record includes the submitted image URL and byte metadata, detected query, market, canonical result URL, extraction method, page counts, timestamps, and nested search metadata. Relative or unavailable dates remain in publishedAtRaw; the Actor does not invent an absolute date. Visual-match source-page links are included only when the result card exposes a reliable association.

The visual-match module is surface-dependent. Source-page records remain available when Bing's headless result page does not render that module.

A live output record is kept at samples/sample-record.json.

Validation

npm test
apify run --purge --input-file INPUT.json
npm run audit:dataset