Job Listings Data by Market: Google Jobs, Remote Boards, France and the Gulf
Which job scraper to use for which hiring market: Google Jobs as the aggregator, Himalayas and HiringCafe for remote and employer jobs, HelloWork for France and Naukrigulf for the Gulf, with inputs, outputs and exact prices.
The actor referenced in this article. Pay only for results delivered.
A hiring dataset is only as good as the board it comes from. A recruiter filling roles in Dubai gets little from a US aggregator, a remote job alert built on a local board misses who can actually apply, and a French salary benchmark needs French contract types. Most teams end up stitching together whatever source they found first, then spend more time cleaning columns than reading the data. This post maps five job scrapers to the hiring market each one covers, so you can pick the right source before you build the pipeline.
Try it live: Google Jobs Scraper. Pay per result delivered.
Which actor for which market
| Actor | Market it covers | You give it | Salary | Gold per 1,000 jobs |
|---|---|---|---|---|
| Google Jobs Scraper | 27 tested countries, aggregated from many boards | Job title and city, or a Google link | Numbers when Google shows one | $3.15 |
| Himalayas Jobs Scraper | Remote jobs worldwide | Keyword, category, company, country, timezone | Range, period and a yearly conversion | $0.70 |
| HiringCafe Jobs Scraper | Jobs from employer career sites and ATS, any country | Keyword plus city, ZIP, state or country | Range with a posted or estimated flag | $0.88 |
| HelloWork Jobs Scraper | France, HelloWork and France Travail | Keyword plus city, postal code or region | Min and max parsed from French text | $0.72 |
| Naukrigulf Jobs Scraper | UAE, Saudi Arabia, Qatar, Kuwait, Oman, Bahrain, Egypt | Keyword plus Gulf city or country | Monthly range in local currency | $0.54 |
Start with Google Jobs to see a market from above, then switch to the board that owns that market when you need volume, deeper filters or fields Google does not carry.
Google Jobs: one search across many boards
Google’s jobs results pull postings from LinkedIn, Indeed, Greenhouse, company career pages and many others into one card per job. The actor turns each card into a row with title, company, location, source (the board Google credits), posted_text, posted_days_ago, posted_at, employment_type, the salary split into salary_min, salary_max, salary_currency and salary_period, the description, Google’s qualifications, responsibilities and benefits, and every apply link in apply_options with its domain and a direct flag.
{
"queries": ["head of sales", "vp sales"],
"locations": ["New York, NY", "Austin, TX"],
"datePosted": "week",
"maxItems": 20
}
Each query runs for each location, so this input is four searches. country defaults to us; set it to de, fr, in, ae or any other code in the list for a different market. Google treats its own date filter loosely, so the actor also drops rows older than your datePosted window, free of charge.
The catch is depth. Google shows 10 jobs per search and has no working next page. Above 10, the actor runs related searches on the same words and removes duplicates. In the README’s test, 100 data analyst jobs in Chicago took 15 searches and about 90 seconds. So treat maxItems as a ceiling, not a promise.
Himalayas: remote jobs by who can apply
For remote roles, the useful question is not where the company sits but where the candidate may live. Himalayas answers it in columns: location_restrictions lists the countries a candidate must live in, is_worldwide is true when anyone can apply, and timezone_restrictions holds the accepted UTC offsets, with location_text and timezone_text as one readable line each.
{
"searchQueries": ["product manager", "data analyst"],
"country": "Germany",
"seniority": ["Mid-level", "Senior"],
"employmentTypes": ["Full Time"],
"sortBy": "recent",
"maxItems": 300
}
Salaries come back as salary_min, salary_max, salary_currency and salary_period, plus salary_yearly_min and salary_yearly_max converted to a year, so hourly and annual pay sort together. onlyWithSalary skips jobs with no range, and skipped jobs are not charged. A single search stops at 5,000 jobs, Himalayas’ own limit.
HiringCafe: jobs straight from employer systems
HiringCafe collects postings from company career sites and applicant tracking systems such as Greenhouse, Lever, Workday, Ashby and iCIMS. That makes it the source to use when you want the employer’s own apply page (apply_url) and the system it came from (ats_source) rather than a job board copy.
{
"searchQueries": ["machine learning engineer"],
"location": "United States",
"visaSponsorship": true,
"onlyPostedSalary": true,
"maxItems": 200
}
Rows also carry seniority_level, role_type, requirements_summary, technical_tools, min_years_experience, visa_sponsorship, and company facts such as company_employees, company_founded and company_funding_stage. salary_is_posted separates ranges the employer published from HiringCafe’s estimates. With includeDetails on, each job is opened for description_text, at about 50 jobs a minute; off, 3,000 jobs took 74 seconds in the README’s test.
HelloWork and France Travail: the French market
French hiring data has its own vocabulary: CDI, CDD, interim, alternance, telework modes and salaries written as “Mensuel de 2900.0 Euros à 4500.0 Euros”. The HelloWork actor searches HelloWork by default and France Travail (formerly Pole emploi) too when you add it to sources, with every row in one schema and a source field.
{
"searchQueries": ["infirmier", "aide soignant"],
"location": "Gironde",
"sources": ["hellowork", "francetravail"],
"contractTypes": ["cdi"],
"postedWithin": "30d",
"maxItems": 200
}
Rows hold contract_type, telework, city, postal_code, salary_min, salary_max and salary_period, and with details on, description_text, skills, education, experience and, from France Travail, rome_code and rome_label. salary_is_estimate flags HelloWork’s own estimates. France Travail serves at most 1,000 offers per search, so split a broad search by departement or contract.
Naukrigulf: the Gulf
Gulf recruiting runs on different signals: nationality asked, monthly pay in AED or SAR, and employer contact blocks. Naukrigulf carries all three. Each keyword is searched in each place you list, spelled as the site spells it.
{
"searchQueries": ["erp consultant", "salesforce developer"],
"locations": ["dubai", "abu dhabi", "riyadh"],
"maxItems": 200
}
Rows carry experience_min_years, experience_max_years, skills, posted_at, and with details on, description, salary_min, salary_max, salary_currency, nationality, gender, industry, functional_area, work_place_type and the contact fields such as contact_name, contact_email and contact_website when the employer publishes them. When Naukrigulf does not know a place, it quietly searches every market; the actor spots this, delivers nothing for that search and writes an info row, so you do not pay for jobs in the wrong country.
One Python script for all five
Every job row has a title, a company field, salary numbers and a link, so the outputs merge with a short mapping step. Rows that carry _type are summary, info or error rows and are never charged; skip them.
import csv
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
inputs = {
"google-jobs-scraper": {"queries": ["data analyst"], "locations": ["Chicago, IL"], "datePosted": "week", "maxItems": 50},
"himalayas-jobs-scraper": {"searchQueries": ["data analyst"], "postedWithinDays": 7, "onlyWithSalary": True, "maxItems": 50},
"hiring-cafe-scraper": {"searchQueries": ["data analyst"], "location": "Chicago", "postedWithin": "1w", "maxItems": 50},
"hellowork-jobs-scraper": {"searchQueries": ["data analyst"], "location": "Paris", "postedWithin": "7d", "maxItems": 50},
"naukrigulf-jobs-scraper": {"searchQueries": ["data analyst"], "locations": ["dubai"], "postedWithinDays": "7", "maxItems": 50},
}
rows = []
for slug, run_input in inputs.items():
run = client.actor(f"themineworks/{slug}").call(run_input=run_input)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
if "_type" in item:
continue
rows.append({
"source_actor": slug,
"title": item.get("title"),
"company": item.get("company") or item.get("company_name"),
"place": item.get("location") or item.get("location_text") or item.get("city"),
"salary_min": item.get("salary_min"),
"salary_max": item.get("salary_max"),
"currency": item.get("salary_currency"),
"period": item.get("salary_period"),
"link": item.get("apply_url") or item.get("job_url"),
})
with open("jobs_by_market.csv", "w", newline="", encoding="utf-8") as f:
writer = csv.DictWriter(f, fieldnames=list(rows[0].keys()))
writer.writeheader()
writer.writerows(rows)
print(f"{len(rows)} jobs written")
Note how each actor spells “last week” differently: "week", 7, "1w", "7d" and "7". Those are the values each input schema accepts, so copy them as they are. Salary periods differ too (Naukrigulf is monthly, Himalayas can be hourly, weekly, monthly or annual), so compare pay within one period before you rank it.
Prices, exactly
All five are pay per event: a price per job delivered plus a flat run-start fee.
| Actor | Free plan per job | Gold and above per job | Run-start fee | When the fee is charged |
|---|---|---|---|---|
| Google Jobs | $0.0081 | $0.00315 | $0.003 | After the first job row reached the dataset |
| Himalayas | $0.0009 | $0.0007 | $0.001 | After a search returned jobs |
| HiringCafe | $0.00112 | $0.00088 | $0.003 | After a search returned jobs |
| HelloWork | $0.0009 | $0.00072 | $0.001 | After a search page returned jobs |
| Naukrigulf | $0.0009 | $0.00054 | $0.001 | After the first call returned jobs |
A run that returns nothing pays no start fee in any of the five. Duplicates found by two searches in the same run, and summary, info and error rows, are never charged. Descriptions and details cost nothing extra on the actors that offer them. As a worked example, 1,000 Naukrigulf jobs on the Gold plan cost $0.54 plus $0.001; the same 1,000 rows from Google Jobs on the Free plan cost $8.10 plus $0.003.
Limits worth knowing
- Google Jobs is shallow per search. Ten jobs per search, and in EU and EEA countries the description is a preview (
description_is_preview: true). Ten countries, including Australia, Ireland and Poland, returned no jobs in testing, and a run there stops with no charge. - Salary coverage varies. Google shows a salary on about 4 of 10 US jobs; about half of HelloWork offers show none; Naukrigulf test runs found a salary on 13 to 63 percent of jobs depending on the role. None of the actors invents a value.
- HiringCafe mixes in remote jobs. A state or country search also returns remote jobs open to that region; filter with
workplaceTypes. - Naukrigulf contact blocks are what the employer chose to publish. Often a website and city, not an email.
- Board terms still apply. Himalayas asks API users to keep the apply link and show Himalayas as the source, and each actor’s legal note puts responsibility for use on you.
If you are building a broader job pipeline, read which job boards allow scraping and how they differ and how to monitor competitor job postings. For LinkedIn as a source, see the LinkedIn Jobs Scraper guide, and for Australia, the SEEK jobs tutorial.
Related
Explore the scraper referenced in this article: inputs, outputs, and pricing, then run it on Apify.
Frequently asked questions
Should I use Google Jobs or a job board scraper? +
Use Google Jobs when you want one search that spans many boards in a country, with salary numbers and every apply link. Use a board scraper when you need depth in one market: Google shows 10 jobs per search, while Himalayas, HiringCafe, HelloWork and Naukrigulf page through thousands per search.
Which of these actors is cheapest per job? +
On the Gold plan, Naukrigulf at $0.54 per 1,000 jobs, then Himalayas at $0.70, HelloWork at $0.72, HiringCafe at $0.88 and Google Jobs at $3.15. On the Apify Free plan, Himalayas, HelloWork and Naukrigulf are $0.90, HiringCafe $1.12 and Google Jobs $8.10.
Am I charged for a search that finds nothing? +
No. All five charge the run-start fee only once jobs come back, and duplicates, summary rows, info rows and error rows are never charged. A run whose searches all return nothing costs nothing.
Can I combine the five outputs into one table? +
Yes. Every job row has a title, a company field, salary_min, salary_max, salary_currency and a link, so a short mapping step merges them. The Python example in this post does that and writes one CSV.
Job Board Scraping 2025: Which Platforms Allow It and How to Do It Right
LinkedIn blocks aggressively, Indeed requires a browser and Naukri needs session warming. The current state of scraping every major job board.
Build a List of Funded Startups That Are Hiring on Wellfound
Turn Wellfound job listings into one row per startup with website, stage, total raised and latest round, using the Wellfound Jobs Scraper.
Compare Salary Bands in Lakhs by City and Experience on Foundit
Build a salary table for one role across Indian cities and experience bands from Foundit.in listings that publish pay, and drop the placeholder figures.
Compare Pay for a Role by Experience Level on Shine
Run one Shine.com search per experience band, keep the listings that show pay, and turn them into salary ranges in lakhs for each band.