Naukri Jobs Scraper

Collect public Naukri job-listing data by keyword and optional location. The Actor returns normalized title, company, location, experience, salary, skills, posting label, public job URL, and any public listing snippet returned by search.

How collection works

1. The Actor calls Naukri's public search endpoint directly with bounded pagination.
2. If Naukri returns an anti-bot response such as HTTP 406, it uses a real browser to load the matching public Naukri search page and captures that page's public search response.
3. The browser fallback starts without a proxy. Only a browser block activates a fresh rotating datacenter proxy session. Metered proxy groups (RESIDENTIAL, GOOGLE_SERP) are never used, and a request for one in proxyConfiguration is replaced with datacenter addresses. Proxy servers you supply yourself in proxyUrls are used exactly as given.

This order keeps ordinary runs inexpensive and never spends metered per-gigabyte bandwidth. Each returned unique job is charged once. Samples, no-result rows, blocked rows, input errors, and network diagnostics are never charged.

Input

• searchQueries: one or more keywords or job titles.
• locations: optional Indian city or region labels. Empty searches nationwide.
• maxItems: maximum unique jobs, 1 to 1,000.
• maxPagesPerSearch: maximum pages per query/location, 1 to 25.
• experience: optional visible experience-label filter.
• includeDetails: retain the public listing snippet returned by search.
• useBrowserFallback: use the browser route only after direct HTTP is blocked.
• maxProxyRetries: maximum proxied browser attempts (rotating datacenter) after a browser block. The legacy key maxResidentialRetries is still accepted as an alias.
• proxyConfiguration: optional. Your own proxyUrls are honoured verbatim; metered Apify groups are replaced with datacenter and the substitution is reported in the log and as an uncharged dataset note.

An empty input intentionally emits one clearly labeled, uncharged sample row so a new task has a non-empty dataset without generating billable data.

Output

jobId, title, company, location, experience, salary, skills, postedAt, description, url, query, searchLocation, and source.

Rows are deduplicated by Naukri job ID across all queries, locations, and pages. Pagination stops at the configured page and item limits, an empty page, or a page with no fresh jobs.

Limitations

• Naukri can require dynamic anti-bot context. A direct endpoint response of HTTP 406 is expected in some environments; the browser fallback addresses this but is slower.
• The proxied browser fallback can still be challenged by Naukri. A blocked run emits an uncharged diagnostic rather than partial or fabricated jobs.
• This Actor collects only publicly accessible job data. It does not log in, apply to jobs, solve CAPTCHAs, or access candidate data.
• Search cards expose the fields Naukri makes public at runtime; listing-detail fields are not guaranteed.

Live pricing assessment

The closest qualifying market floor found in the July 25, 2026 scan was $0.0004/job. Apify cloud returned HTTP 406 on direct search, then failed to capture a public search response in either direct-browser or proxied-browser mode. A prior India-residential attempt cost $0.0170047 and returned zero jobs, so residential was removed outright on August 15, 2026: it bought nothing and billed the owner per gigabyte. An August 15, 2026 side-by-side probe of the search endpoint returned the same HTTP 406 recaptcha required from a plain address, from Apify datacenter, and from Apify residential India. This Actor therefore remains private and unpriced; blocked runs settle successfully with an uncharged diagnostic.