Naukri Scraper

Scrape Naukri.com, India's largest job board, by keyword or by link. Every record carries the title, company, skills, experience, salary, location and posting date straight from the search results; with details on, each record also carries the full description, key skills, education, employment type, workplace label, industry, company details and view/apply counts from the job's own page.

Why a browser session

The search results are rendered by a client-side app that reads a same-origin JSON interface guarded by a short-lived token generated inside the page. A plain HTTP client is refused (HTTP 406) even with a valid cookie jar, so the actor drives one persistent browser session for the whole run: it loads the first page of each search, harvests the payload the app itself fetched, then reads later pages through the page's own connection. One session per run, no per-page churn.

Modes

• Search: one or more keywords, optionally with one location. Each keyword's results are walked forward page by page until they end or a cap stops the run.
• URL: paste links instead.
  • search pages (https://www.naukri.com/python-developer-jobs-in-bangalore) keep every filter you applied on the site, because the page path carries them; they are walked like a search.
  • single job pages (https://www.naukri.com/job-listings-...-<jobId>) are read directly.

Output fields

Each dataset item includes: jobId, title, url, companyName, companyId, companyLogo, skills, experience, minimumExperience, maximumExperience, salary, location, descriptionSnippet, vacancy, postedAt, postedDaysAgoLabel, currency, walkinJob, searchScope. With Fetch full job details on it also carries description, locations, education, employmentType, jobType, industry, functionalArea, jobRole, workFromHomeType, viewCount, applyCount, companyDetails, companyRating and the salary breakdown in salaryDetail.

Limits and cost

• maxItems (default 20) is the total record cap across all keywords or links; set 0 for unlimited.
• maxPages (default 0) is unlimited; maxItems is the one cap that stops a run.
• Fetch full job details (default on) reads one job page per record. It is charged once per record whose job page read actually succeeded. Turn it off for the cheap card-only surface.

Scheduled runs, incremental mode

Turn on Incremental changes for scheduled runs and run the same search daily: the first run returns everything as NEW, later runs return only NEW, UPDATED and REAPPEARED records. Records that vanished are marked EXPIRED once a run has scanned the whole search (a run that was capped, resumed, or served fewer results than the site reported skips expiry detection rather than guess). resumeFromRunId continues one specific interrupted run.

Connection

The actor runs through the Apify proxy on the standard shared pool by default. Residential proxies in India are supported if your runs are refused, at the usual per-GB premium.

Send results into your apps (MCP connectors)

Optionally push results straight into another app when the run finishes. Add a connector in the input under mcpConnectors, set notionParentPageUrl when the connector is Notion, and cap the export with maxNotifyListings. Leave the list empty to only write the dataset.

Input parameters

Parameter	Type	Description
mode	select	search or url.
keywords	list	Job keywords for search mode.
location	string	Optional city applied to every keyword.
urls	list	Job or search page links for URL mode.
fetchDetails	boolean	Read each job's page for the full detail set (per-job surcharge).
maxItems	integer	Total record cap, 0 = unlimited.
maxPages	integer	Pages per search, 0 = unlimited.
proxyConfiguration	object	Proxy settings; standard shared pool by default.
resumeFromRunId	string	Continue a previous run or dataset.
incrementalMode	boolean	Remember state between scheduled runs, return only changes.
stateKey	string	Name or share an incremental monitoring campaign.
emitUnchanged	boolean	Also return UNCHANGED rows (billed).
emitExpired	boolean	Also return EXPIRED rows once a full scan allows it.
mcpConnectors	list	Connector IDs for the export side-channel.
notionParentPageUrl	string	Parent page for the Notion connector.
maxNotifyListings	integer	Cap on exported records per run.

Usage notes

• Keyword searches match the site's own behaviour closely: nonsense keywords return few or no results, and the run then finishes with an honest "matched nothing" outcome.
• Job boards churn: postings move, expire and re-appear between runs. In incremental mode a re-appearing posting is REAPPEARED, not NEW, and its firstSeenAt stays from the first sighting.
• All URLs in the dataset point back to the live job page.

Limitations

• Naukrigulf (the Gulf sister site) is not covered by this actor; naukri.com India search results do include some Gulf postings where the site serves them.
• The site serves 20 results per page; the actor respects that page size.
• Ages shown on cards ("14 Days Ago") are volatile display labels; the exact posting timestamp in postedAt is the reliable field.