Generate your queries

Four searches from one role and city. Copy each into Google — they surface different things.

Generate Google and X-ray searches
Careers pages
("data analyst" OR "business analyst" OR "analytics analyst") "Bengaluru" ("careers" OR "we're hiring") -senior
Applicant tracking systems
("data analyst" OR "business analyst" OR "analytics analyst") "Bengaluru" -senior
(site:boards.greenhouse.io OR site:jobs.lever.co OR site:jobs.ashbyhq.com OR site:apply.workable.com OR site:jobs.smartrecruiters.com OR site:*.myworkdayjobs.com)
Hiring posts, not listings
("we're hiring" OR "my team is hiring" OR "looking for")
("data analyst" OR "business analyst" OR "analytics analyst") "Bengaluru"
One company only
site:company.com ("data analyst" OR "business analyst" OR "analytics analyst") "Bengaluru"

Results depend on what has been indexed, so none of these is exhaustive. They surface listings the aggregators never picked up, which is the point.

What “X-ray” actually means

It is a recruiter’s term for using a general search engine to look *inside* a site that has its own weak search — or no public search at all. Applicant tracking systems are the classic case. Thousands of companies host their listings on Greenhouse, Lever, Ashby, Workable and Workday, each on a predictable URL, and none of those platforms offers a useful cross-company search. Google has indexed all of them.

So instead of asking a job board what exists, you ask Google to show you every Greenhouse page mentioning your role. The results are current, unsyndicated, and posted by the employer rather than a reseller.

The operators worth knowing

site:boards.greenhouse.io
Restrict results to one host. The single most useful operator for job hunting, because it turns every ATS into a searchable job board.
"exact phrase"
Match the words together and in order. Essential for job titles, which otherwise match any page containing the words separately.
-word
Exclude. Google’s equivalent of NOT. Use sparingly — it removes whole pages, not just the phrase.
OR
Either term. Must be capitalised. This is how you cover a title family in one query.
intitle:
The term must appear in the page title. Useful for filtering to actual listing pages rather than blog posts mentioning a role.
after:2026-09-01
Only pages Google believes are newer than that date. Approximate, but a decent freshness filter when the Tools menu is not enough.

The hosts worth searching directly

These carry the bulk of listings at technology and startup employers. Substitute your role into each.

SystemHost to searchTypically used by
Greenhousesite:boards.greenhouse.ioMid-size and large tech companies, well-funded startups
Leversite:jobs.lever.coStartups and scale-ups
Ashbysite:jobs.ashbyhq.comNewer, fast-growing startups
Workablesite:apply.workable.comSmall and mid-size companies across sectors
SmartRecruiterssite:jobs.smartrecruiters.comLarge enterprises, retail, services
Workdaysite:*.myworkdayjobs.comEnterprises, banks, consultancies
Zoho Recruitsite:*.zohorecruit.comCommon with Indian SMEs and agencies

Queries to keep

Paste, swap the role and city, and run them weekly. The last two find things no job board will ever show you.

Google queries
Every Greenhouse listing for your role
site:boards.greenhouse.io ("data analyst" OR "business analyst") "Bengaluru"

Across four systems at once
("product manager" OR "APM") "remote"
(site:boards.greenhouse.io OR site:jobs.lever.co OR site:jobs.ashbyhq.com OR site:apply.workable.com)

Careers pages, not aggregators
("frontend engineer" OR "UI engineer") "Pune" ("careers" OR "we are hiring") -site:naukri.com -site:indeed.com

Hiring posts rather than listings
("we're hiring" OR "my team is hiring" OR "DM me your CV") ("backend engineer" OR "SDE") "Hyderabad"

One company, everything open
site:stripe.com/jobs ("engineer" OR "designer")

Searching for hiring language, not job titles

A large share of hiring never becomes a formal listing. A team lead posts that they are looking for someone, a founder mentions it in a newsletter, a company adds a line to an about page. None of that is indexed as a job, and none of it appears on Naukri, Indeed or LinkedIn’s job tab — but all of it is text on the open web, and text is what a search engine is for.

Searching phrases like "we’re hiring", "join our team" or "send your CV to" alongside your role surfaces this layer. The competition on it is a fraction of what it is on a job board, because reaching it requires knowing it exists.

Honestly, where this technique fails

It is excellent for
  • Technology, startups and any employer using a modern ATS
  • Finding roles days before they reach an aggregator
  • Watching a specific shortlist of companies you want to work at
  • Roles at companies too small to pay for job board listings
It is poor for
  • Employers who only post to a national board and have no careers page
  • Sectors where hiring runs through agencies — search results are then mostly agency pages
  • Knowing whether a listing is still open, since indexing lags reality
  • Volume: it returns fewer results than a board, and they need more judgement

A weekly twenty-minute routine

  1. Run the four generated queries. Role family, ATS hosts, careers pages, hiring language. Set the date filter to the past week on each.
  2. Open everything plausible in tabs. Judge later. Stopping to evaluate each result is how twenty minutes becomes ninety.
  3. Save the ones worth applying to. These pages disappear without notice and are much harder to find a second time than a board listing.
  4. Add newly-found companies to a watch list. A company hiring one role you want is likely to hire another in three months. site: their careers page directly next time.
  5. Apply in a separate session. Sourcing and applying are different kinds of work and mixing them halves the output of both.

The real cost of doing this properly

Four queries a week across two roles and three cities is twenty-four searches, each returning results that need reading and judging. Done properly it is the highest-yield sourcing there is. Done at the end of a working day it is the first thing to get skipped, and it is skipped by nearly everyone — which is precisely why the competition on these listings is so much lower than on the boards.

Questions

What is X-ray search for jobs?

Using a general search engine to look inside sites that have weak or no public search of their own — most usefully applicant tracking systems such as Greenhouse, Lever and Workday, where thousands of companies publish listings that never reach a job board.

How do I search Google for jobs on company career pages?

Combine your role in quotes with site: restricted to the host. For example: site:boards.greenhouse.io ("data analyst" OR "business analyst") "Bengaluru". Then set Tools → Past week to filter out stale indexed pages.

What Google operators are useful for job searching?

site: to restrict to one host, quotation marks for exact job titles, OR for alternative titles, a minus sign to exclude, intitle: to require a word in the page title, and after: for a rough date filter. The Tools menu date filter is more reliable than after:.

How do I find jobs on Greenhouse or Lever?

Search site:boards.greenhouse.io or site:jobs.lever.co followed by your role in quotes and your city. Neither platform offers a cross-company job search, so a search engine is the only practical way to read across all of them at once.

Is Google Jobs the same as X-ray searching?

No. Google Jobs is a curated listings panel built from structured data that employers and boards submit. X-ray searching queries the open index directly, which reaches pages that were never submitted to any jobs feed.

Why do X-ray searches return closed jobs?

Search indexes lag reality — a page can stay indexed for weeks after a role is filled, and many ATS pages simply return an error once pulled. Filtering to the past week under Tools removes most of the stale results.

Does X-ray search work for non-tech jobs?

Partly. It works wherever employers publish their own listings, which increasingly includes retail, healthcare and financial services on systems like Workday and SmartRecruiters. It works poorly in sectors where hiring runs mainly through agencies, since results then skew towards agency pages rather than employers.

Read next

LinkedIn Boolean search: a generator, and how the operators actually workBuild a Boolean string for LinkedIn job search in a few clicks, then learn the four operators behind it — including the two that quietly exclude roles you wanted.Where to find jobs beyond LinkedInLinkedIn is one channel of several, and rarely the first place a role appears. A guide to job boards, applicant tracking systems, company careers pages and the channels that get the fewest applicants.How to find jobs posted in the last 24 hoursReply rates fall sharply with the age of a listing, and “posted 2 days ago” is often a lie. How to filter for genuinely new roles on every major board, and how to spot a recycled posting.