Google Jobs Scraper

Google Jobs Scraper: Deep Search, Salaries & Clean Apply Links

By Bight Systems. Turn any Google Jobs search into a clean, spreadsheet-ready dataset: job title, company, location, remote/hybrid flag, employment type, salary parsed into numbers, posted date as a real date, full description, qualifications / responsibilities / benefits, and apply links with mismatches filtered out.

Google Jobs collects listings from LinkedIn, Indeed, Glassdoor, ZipRecruiter, company career sites and hundreds of local boards, so one run gives you one de-duplicated feed covering all of them.

Last verified: November 10, 2026, on 39 live searches in 6 countries (US, UK, Canada, Germany, India, France), plus single-request checks in 30 more countries (36 countries with live results in total; see Supported countries).

---

Why this scraper

	What you get
More than 10 jobs per search	Google shows only ~10 jobs per search. Standard depth re-runs your search with Google's own filter phrases (full time, part time, posted this week…) and merges the results without duplicates: median 55 jobs per search in our validation. Deep depth: median 101.
Free during launch, then pay per job	No Actor fee during the launch period. Later pricing will be per job returned, not per page: empty searches, duplicate jobs and unsupported countries won't cost anything in results.
Salary as numbers	$120K–$150K a year, £14–£16 an hour, 55.000 € bis 75.000 € pro Jahr, ₹5L–₹8L a year, 32 k € à 38 k € par an, R$ 5,5 mil a R$ 16,5 mil por mês → min, max, ISO currency, period, plus an annualized range for comparisons.
Real dates	Google only shows "3 days ago". We also give postedDate (ISO date) and postedHoursAgo, computed from the run time.
Clean apply links	Google merges similar postings, so a job's link list can include links to other jobs. We drop links whose URL clearly names a different job, label every link (match / employer / unverified), and pick the most trustworthy primary apply link. Optional live check of the primary link.
New-job alerts	"Only new jobs" mode remembers what earlier runs returned and outputs new jobs only. Schedule it daily for alerts.
Honest failures	Retries with backoff, timeouts, partial results kept if Google misbehaves, a run summary (OUTPUT) explaining every search, and a clear message instead of an empty "success" when a country is not supported.

Use cases

• Recruiting & staffing agencies: find companies hiring for the roles you place, with apply links and descriptions.
• Job boards & aggregators: a daily feed of fresh listings per category and city.
• Salary benchmarking: normalized pay ranges by title, city and country.
• Sales intelligence: hiring activity as a buying signal for your product.
• Labor-market research & AI agents: structured, de-duplicated job data through the API.

Quick start (60 seconds)

1. Enter one or more Search queries (e.g. software engineer), or paste Google Jobs URLs straight from your browser.
2. Optionally add Locations (e.g. Austin, TX) and pick a Country.
3. Choose Search depth (Standard is a good default) and click Start.
4. Open the Overview tab of the dataset, or export to CSV, Excel, JSON, or Google Sheets.

Example input

{
  "queries": ["software engineer", "registered nurse"],
  "locations": ["Austin, TX"],
  "country": "US",
  "depth": "standard",
  "datePosted": "week",
  "maxJobsPerQuery": 100,
  "onlyNewJobs": true,
  "stateKey": "austin-tech-nursing"
}

Input fields

Field	What it does	Default
queries	Job titles/keywords, one per search. You can include a place ("nurse Chicago IL").	required unless startUrls
startUrls	Google Jobs search URLs (Jobs tab udm=8, or older ibp=htl;jobs links). We read the search text, country (gl or the domain), language (hl) and date/employment-type chips. Depth, filters and output settings still apply.	none
locations	Cities/regions; every query × every location is searched.	none
country	Google Jobs market (domain, region, currency).	US
language	Interface language; auto = the country's main language.	auto
datePosted	any, today, 3days, week, month.	any
remoteFilter	any, remoteOnly, excludeRemote.	any
employmentType	One type via Google's own filter: any, FULLTIME, PARTTIME, CONTRACTOR, INTERN, TEMPORARY.	any
employmentTypes	Several types at once (e.g. ["PARTTIME","CONTRACTOR"]). Each type is searched separately and merged without duplicates; jobs Google labels with another type are removed (jobs with no type shown are kept).	none
postedWithinHours	Keep only jobs posted at most X hours ago (e.g. 12, 48). Also sets the matching Google date window. Jobs without a posted date are removed.	none
salaryOnly	Keep only jobs with a parsed salary.	off
minAnnualSalary	Keep jobs whose annualized salary (top of range) is at least this amount, in the job's currency. Jobs without a salary are removed.	none
includeCompanies / excludeCompanies	Keep / drop jobs whose company name contains one of these words (case-insensitive).	none
depth	single (1 request, ~10 jobs), standard (≤9 requests, ~50), deep (≤21 requests, ~100+).	standard
maxJobsPerQuery / maxJobsTotal	Caps per search and per run (0 = no run cap).	100 / 0
onlyNewJobs + stateKey	Skip jobs returned by earlier runs with the same key (remembered 90 days).	off / default
includeDescription, includeDescriptionHtml, includeHighlights	Control the large text fields.	on, off, on
outputFormat	full (arrays, best for API), compact (no long text), flat (lists joined into text for CSV/Sheets).	full
applyLinkFilter	standard (drop clear mismatches), strict (only confirmed links), off.	standard
verifyPrimaryApplyLink	Open the primary link and confirm the page title matches. Slower.	off
stripTrackingParams, dedupeAcrossQueries	Remove utm_* parameters; return a job once across searches.	on, on
maxConcurrency, requestTimeoutSecs, maxRetries	Advanced reliability settings.	10, 60, 3
healthCheck, healthThresholds, healthLinkSample	Quality-monitoring mode for scheduled runs (see FAQ).	off, defaults, 40

Output

Each dataset item is one job. Example (shortened):

{
  "title": "Staff Accountant (Hybrid)",
  "company": "Northwest Administrators, Inc.",
  "location": "Seattle, WA",
  "isRemote": false,
  "isHybrid": true,
  "workArrangement": "HYBRID",
  "employmentType": "Full-time",
  "employmentTypes": ["FULL_TIME"],
  "salaryText": "$30.70–$41.54 an hour",
  "salaryMin": 30.7,
  "salaryMax": 41.54,
  "salaryCurrency": "USD",
  "salaryPeriod": "HOUR",
  "salaryAnnualMin": 63856,
  "salaryAnnualMax": 86403,
  "postedAtText": "5 days ago",
  "postedDate": "2026-11-05",
  "postedHoursAgo": 120,
  "sourceBoard": "Indeed",
  "companyWebsite": null,
  "primaryApplyUrl": "https://builtin.com/job/staff-accountant-hybrid/12098681",
  "primaryApplySite": "Built In",
  "primaryApplyLinkStatus": "confirmed",
  "applyLinks": [
    {"url": "https://www.indeed.com/viewjob?jk=a5149a544102405f", "site": "Indeed", "isDirect": true, "match": "unverified"},
    {"url": "https://builtin.com/job/staff-accountant-hybrid/12098681", "site": "Built In", "isDirect": false, "match": "match"}
  ],
  "applyLinksCount": 4,
  "applyLinksDropped": 4,
  "description": "Are you a Staff Accountant who enjoys maintaining financial records…",
  "qualifications": ["Strong understanding of accounting principles and practices", "…"],
  "responsibilities": ["Create journal entries to record transactions on a modified cash basis", "…"],
  "benefits": ["…"],
  "googleJobId": "zfGMTblnYdlpkpevAAAAAA==",
  "jobHash": "ba4fbf897bb02f9e",
  "mergedPostingsCount": 8,
  "googleJobsUrl": "https://www.google.com/search?ibp=htl;jobs&q=accountant+jobs+Seattle,+WA&htidocid=…",
  "searchQuery": "accountant",
  "searchLocation": "Seattle, WA",
  "foundViaSlice": "base",
  "country": "US",
  "language": "en",
  "scrapedAt": "2026-11-10T13:06:27Z"
}

Field reference

Field	Meaning
title, company, location	As shown on Google Jobs.
isRemote, isHybrid, workArrangement	Remote = location "Anywhere" or remote in the title/labels; hybrid = "hybrid" in title or description. REMOTE / HYBRID / ON_SITE.
employmentType, employmentTypes	Text as shown, plus normalized list (FULL_TIME, PART_TIME, CONTRACTOR, INTERN, TEMPORARY, …).
salaryText, salaryMin, salaryMax, salaryCurrency, salaryPeriod	Pay as shown and parsed. Currency is ISO 4217; a bare $ uses the country's dollar (USD in the US, CAD in Canada). Period: HOUR, DAY, WEEK, MONTH, YEAR.
salaryAnnualMin, salaryAnnualMax	Annualized (hour × 2080, day × 260, week × 52, month × 12) for comparisons.
postedAtText, postedDate, postedHoursAgo	Google's relative text, and the date computed from the run time (approximate by nature; "30+ days ago" is a lower bound).
sourceBoard	The board Google credits ("via LinkedIn").
companyWebsite	Employer domain (e.g. goldmansachs.com) when an apply link or a link in the description is on a domain that matches the company name; job boards and ATS hosts are never used. Filled for roughly 15% of jobs in our tests; null otherwise.
primaryApplyUrl, primaryApplySite, primaryApplyLinkStatus	Best apply link. Status: confirmed (live page title checked), employer (company's own site), match (URL names this job), unverified (URL has no readable job name, e.g. an ID-only Indeed link).
applyLinks[]	Every kept link: url, site, isDirect (Google's "Apply directly" label or an applicant-tracking/career site), match.
applyLinksCount, applyLinksDropped	Links kept and links removed by the filter.
description, descriptionHtml	Full description text (and markup if requested).
qualifications, responsibilities, benefits, otherHighlights	Google's highlight bullets, when the job has them.
degreeNote	e.g. "No degree mentioned", when Google shows it.
googleJobId, jobHash	Google's job id, and our stable hash of title + company + location (use it to de-duplicate across runs).
mergedPostingsCount	How many postings Google grouped into this job.
googleJobsUrl	Link to the job on Google.
searchQuery, searchLocation, searchText, foundViaSlice	Which search found the job, the exact text sent to Google, and which depth slice found it.
country, language, scrapedAt	Run context; scrapedAt is UTC.

The Overview, Salaries, Apply links and Descriptions & highlights tabs in the dataset give ready-made tables. Each run also writes a summary to the key-value store record OUTPUT: per-search request counts, jobs found, failures, skipped searches and notes.

What our validation measured (Nov 10, 2026)

39 live searches (US, UK, Canada, Germany, India, France; cities, remote, niche titles, filters):

Metric	Result
Jobs per search, Standard depth	median 55 (32 searches)
Jobs per search, Deep depth	median 101 (83–116)
Searches with 10+ jobs	38 of 39 (the one miss was a temporary empty page from Google; the search returned 55 jobs on re-run, and an automatic fallback for this case has since been added)
Requests that failed after retries	3 of 372 (0.8%); runs kept everything else
Title / company / description filled	100% / 100% / 99%
Employment type / posted date / salary present	97% / 76% / 43% (only as often as Google shows them)
Salary text parsed into numbers	99.9% of salaries shown
Primary apply link correct (live page check, 50 sampled jobs)	26 of 26 readable pages matched the job; 24 sites block automated checks
Links our filter dropped that really were another job	29 of 29 readable pages (40 dropped links sampled)
Run time	about 20–105 s per search (Standard/Deep); searches in one run go in parallel

Limits we found and tell you about: some secondary apply links that we keep (unverified) still point to similar jobs, so use primaryApplyUrl or applyLinkFilter: "strict" when link accuracy matters most. Highlights exist only on some jobs (about 30–45%).

Pricing

Free during the launch period. There is no Actor fee right now. You only pay your normal Apify platform usage (compute and Google SERP proxy), which came to about $0.50 per 1,000 jobs in our tests.

After launch, the Actor will switch to pay-per-result pricing of about $2.50 per 1,000 jobs, with lower prices on higher Apify plans. We will announce the change in advance in the Changelog, and Apify will show it on the Pricing tab before it takes effect.

How pricing will work:

• You pay only for jobs returned. There is no per-page fee and no start fee.
• Empty searches, duplicate jobs, jobs skipped by "Only new jobs" mode, and unsupported countries are never charged as results.
• Set Maximum cost per run in Apify to cap spend. The run stops cleanly and keeps the results so far.

How it compares with typical Google Jobs results

	Typical Google Jobs scraper	This Actor
Jobs per search	~10 (one Google page); more pages cost more	~50 (Standard) or ~100+ (Deep) per search, de-duplicated
Price model	Per page of ~10 results ($3.50–$9 per 1,000 jobs), or per job plus platform usage	Free during launch, then per job (about $2.50 per 1,000)
Salary	Raw text	Text plus min, max, ISO currency, period and an annualized range
Posted date	"3 days ago"	Text plus ISO date and hours ago
Apply links	Everything Google lists, including links to other merged jobs	Mismatches dropped, every link labelled, one ranked primary link
Unsupported countries	Sometimes empty "successful" runs	Clear stop message, no result charges
Repeat monitoring	Often re-pay for the same jobs	"Only new jobs" mode
Quality monitoring	Rarely offered	Built-in health-check mode (see FAQ)

Supported countries

Status	Countries
Tested Nov 10, 2026 (Standard and Deep depth)	United States, United Kingdom, Canada, Germany, India, France
Tested Nov 10, 2026 (single-request check: 8–10 jobs each, title/company/location/description/apply link filled on every job)	Netherlands, Spain, Brazil, Mexico, South Africa, Austria, Switzerland, Belgium, Italy, Portugal, Argentina, Colombia, Peru, Nigeria, Kenya, Ghana, Egypt, UAE, Saudi Arabia, Pakistan, Bangladesh, Singapore, Malaysia, Philippines, Indonesia, Japan, Hong Kong, Denmark
Works, with limits (tested Nov 10, 2026)	Greece: results are intermittent (Greek-language searches sometimes come back empty; English searches such as "accountant Athens" work). Chile: Google shows very few listings (2 jobs for "ingeniero Santiago").
Not supported by Google Jobs (run stops immediately with a clear message and no result charges)	Australia, New Zealand, Ireland, Poland and Sweden (checked Nov 10, 2026), plus South Korea, Czechia, Turkey, Ukraine, Israel, Norway, Finland, Estonia, Latvia and Lithuania (per public reports)

Salaries, posted dates and employment types are parsed in English, German, French, Spanish, Portuguese, Dutch, Italian, Japanese, Greek, Danish and Indonesian formats (for example € 3.171–€ 4.776 per maand, R$ 5,5 mil a R$ 16,5 mil por mês, 年 ￥450万～￥650万, Rp 3,5 jt–Rp 4 jt per bulan). For depth slices, Standard/Deep use translated filter phrases for English, German, French and Spanish, and English phrases elsewhere. Searches whose location names an unsupported country are skipped the same way. If a configured country doesn't work for you, please open an issue.

Filters you control (not charged)

employmentTypes, postedWithinHours, salaryOnly, minAnnualSalary, includeCompanies and excludeCompanies are applied after Google returns results. Removed jobs are never pushed to the dataset, so they never count as results. The run summary (OUTPUT record) shows how many were removed and why in filteredOut (e.g. {"noSalary": 26, "salaryBelowMinimum": 5}). Tip: combine filters with Standard or Deep depth, because filtering happens after the search.

Use from AI agents (MCP)

AI agents such as Claude, Cursor, ChatGPT-compatible clients or your own LangChain/LangGraph agent can call this Actor through the Apify MCP server.

• Hosted (recommended): add the MCP server URL https://mcp.apify.com?tools=bightsystems/google-jobs-scraper to your client. It signs in with OAuth, or send the header Authorization: Bearer <YOUR_APIFY_TOKEN>. The tools= part loads only this Actor, which keeps the agent's tool list small.
• Local: npx -y @apify/actors-mcp-server --tools bightsystems/google-jobs-scraper with APIFY_TOKEN set in the environment.

Example client config:

{
  "mcpServers": {
    "google-jobs": {
      "url": "https://mcp.apify.com?tools=bightsystems/google-jobs-scraper",
      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }
    }
  }
}

Tips for agents: call the Actor with a small input (one query, one location, "depth": "single" or "standard", "outputFormat": "compact"), then read the results with the dataset tool. Use maxJobsTotal to cap the job count and callOptions such as maxItems / maxTotalChargeUsd to cap spend. The OUTPUT record has a short run summary an agent can read first. Example prompt: "Find part-time or contract accountant jobs in Chicago posted in the last 48 hours that pay at least $60,000, and list company, salary and apply link."

Integrations

• API: POST https://api.apify.com/v2/acts/bightsystems~google-jobs-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN with the input JSON returns the jobs directly. Python and JavaScript clients: apify-client.
• Make, Zapier, n8n: use the Apify app/node ("Run Actor" then "Get dataset items"), or trigger a scheduled run and send new jobs to Google Sheets, Airtable, Slack or your CRM.
• Schedules: Apify Schedules + onlyNewJobs: true = a daily alert that returns only new jobs.
• Exports: JSON, CSV, Excel, XML, RSS, or Google Sheets from the dataset. Use outputFormat: "flat" for one clean row per job in spreadsheets.

FAQ

Why do I get fewer jobs than I expected? Google itself only has a limited number of listings per search; narrow searches (niche titles, small towns) can return 10–30 jobs even at Deep depth. Try a broader title or location, or Deep depth.

Why did a remote search change my query? Remote searches without a location return few or no jobs on Google, so we add the country name (e.g. "UX designer remote jobs United States"). The run summary notes this.

Why are some apply links missing compared to Google? The filter dropped links whose URL names a different job. Set applyLinkFilter to off to get everything, each labelled.

Is postedDate exact? It is computed from Google's relative text ("3 days ago") at run time, so it is accurate to the day for recent jobs and a lower bound for "30+ days ago".

My run returned 0 jobs. Check the status message and the OUTPUT record. Common causes: an unsupported country, a misspelled query, or a very narrow location.

Do you scrape LinkedIn or Indeed directly? No. We read Google Jobs results, which list jobs from those sites, and return the links Google provides.

Is this legal? We collect publicly shown job listings and do not collect personal data beyond what employers publish in job posts. You are responsible for using the data in line with the laws that apply to you (e.g. GDPR) and with the terms of the sites you link to.

Can I monitor data quality myself? Yes. Turn on healthCheck. After the searches, the run scores itself: jobs per search, field fill rates, failed requests, and a live check of a sample of primary apply links. It is marked FAILED, with the reasons in the status message, if a threshold is missed (thresholds can be changed in healthThresholds). Combine it with an Apify Schedule and run-failure notifications.

Can I paste a link from Google? Yes. Search on Google, open the Jobs tab, copy the address bar into startUrls. Links from other sites are skipped with a note in the run summary. If you only want the URLs, clear the example queries.

Is there a search radius? Not yet. Google currently ignores the radius parameter on its Jobs tab in our tests (5 km and 300 km returned the same town-only results), so we haven't added a radius setting that wouldn't work. Put the city you want in locations, or add nearby cities as extra locations.

Something broke or a field is wrong. Open an issue on the Actor's Issues tab with the run ID and we will look into it.

Changelog

See the Changelog tab. Field changes are announced there.