Naukri Role Scraper — compare Indian roles by salary

Give two or more Naukri searches and get back one row per role or city, ranked by median salary — with a confidence range, a disclosure rate, and a number most job-board scrapers never show you: how much of the "demand" comes from a single poster.

This is the comparison companion to Naukri Jobs Scraper. That one returns one row per job ("what is being advertised"). This one returns one row per search ("which role pays more, and is the demand real").

What you get per role

Field	Example
role	data analyst in bangalore
medianSalaryInr	1200000
medianSalaryInrRange	[950000, 1450000] — 95% range for the median
medianSalaryUsd, inrPerUsdUsed	13636, 88 — the rate used is always returned
jobs, jobsWithSalary, salaryDisclosureRate	40, 11, 0.275
medianExperienceMinYears	3 — the entry bar for this role
uniqueCompanies, topPosterShare, topPoster	18, 0.35, TalentBridge Staffing
medianCompanyRating, medianPostedDaysAgo	3.9, 2
rank, notRankedReason	1, or why the row was left out

Why topPosterShare is on every row

On Indian job boards a single staffing firm often posts the same requirement dozens of times. Count the listings and you will read that as demand. This Actor reports the share held by the largest poster, and the summary names any role where one company holds 30% or more — so a large job count can be discounted when it deserves to be.

What this Actor refuses to do

Rank on a point estimate alone. Medians from small samples swing hard. Every row carries a range, and if the top two ranges overlap the summary says topTwoTooClose: true — the data cannot tell you which is higher. Roles with fewer than 6 salary-bearing jobs get rank: null.

Hide the denominator. Naukri listings very often say "Not disclosed". salaryDisclosureRate is on every row, because a median drawn from a quarter of the sample is not the same claim as one drawn from all of it.

Invent an exchange rate. The USD column uses the fixed rate you pass in, and that rate is returned on every row. Leave it out and no USD figure is produced.

Summary (key-value store, BENCHMARK_SUMMARY)

roles, ranked, bestPaying, bestPayingMedianInr, topTwoTooClose, medianSalaryDisclosureRate, concentrated, notRanked.

Input

{
  "queries": ["data analyst", "data engineer", "data scientist"],
  "samplePerQuery": 40,
  "inrPerUsd": 88
}

You can pass plain search terms or Naukri listing URLs. A single query is rejected — one search is not a comparison. You are charged per role, not per job.

Limits, stated plainly

• Salary figures come from what employers chose to publish. They are not payroll data.
• Only job listings are read. Nothing about applicants is collected.

Unofficial; not affiliated with Naukri or Info Edge.