Trabalha Brasil Scraper: Brazilian job vacancies

Scrape Trabalha Brasil (trabalhabrasil.com.br), a Brazilian job board, by role and city (vagas de emprego): title, employer, city and state, posted pay in BRL with its period, home office or on site, contract, number of vacancies, posting date and the full description. Export to CSV, Excel or JSON, or use the API.

USD 0.59 per 1,000 jobs. No login, no browser.

What you get

• Only the pay the ad shows. Most Trabalha Brasil ads say "Salário: A combinar" (to be agreed) while the page's structured data still carries a pay range: the site's estimate for the role, the same on every nurse ad we read. That estimate never goes into the salary fields; those ads get extra.salaryToBeAgreed. In our 20-job test, 2 ads had a posted pay (BRL per month) and 18 only the estimate.
• The pay period as written: "R$ 10.848,00 por mês" is 10,848 BRL per MONTH, "R$ 800,00 por semana" is per WEEK.
• Home office, on site or hybrid from the ad (3 of 20 home office in our test), the contract as written ("Efetivo/CLT", "Estágio", "Aprendiz"; internships and apprenticeships are also mapped to INTERN and APPRENTICE), and the number of vacancies when the title says "9 Vagas de ..." (extra.vacancies).
• Search by role and city the way the site does (enfermeiro in São Paulo/SP), a role in all of Brazil, every job in a city, or pasted list URLs.

Input examples

{ "keywords": ["motorista", "vendedor"], "city": "Curitiba", "state": "pr", "maxJobs": 200 }

Nurse jobs in São Paulo posted in the last 30 days, new ones only (daily schedule):

{ "keywords": ["enfermeiro"], "city": "São Paulo", "state": "sp", "postedWithinDays": 30, "onlyNewJobs": true }

Example row

From a saved Trabalha Brasil list and job page (empty fields left out, description shortened):

{
  "jobId": "13664910",
  "title": "Enfermeiro",
  "company": "Irmãs Hospitaleiras",
  "location": "São Paulo/SP",
  "city": "São Paulo",
  "region": "SP",
  "country": "BR",
  "remote": false,
  "workplaceType": "onsite",
  "salaryMin": 10848,
  "salaryMax": 10848,
  "salaryCurrency": "BRL",
  "salaryPeriod": "MONTH",
  "salaryText": "R$ 10.848,00 por mês",
  "employmentTypeText": "Efetivo/CLT",
  "category": "Enfermeiro",
  "datePosted": "2026-09-24",
  "descriptionText": "Descrição Geral:Vaga para enfermeiro(a) no caps parelheiros - Irmãs hospitaleiras.\nSalário: R$ 10.848,40\nInsalubridade: r$324,20\nJornada de ...",
  "jobUrl": "https://www.trabalhabrasil.com.br/vagas-de-emprego-em-sao-paulo-sp/enfermeiro/13664910",
  "sourceBoard": "Trabalha Brasil",
  "duplicateKey": "1a9e1dc2aa1a3cfa",
  "language": "pt"
}

Good to know

• Many ads hide the employer: "Empresa Confidencial" gives no company name and extra.confidentialEmployer (8 of 20 ads in our test).
• Old ads stay listed. Some ads still on the site were posted years ago (2023 dates among the first results), and the list is not sorted by date. Use Posted within (days) to drop them; it works from the date on the job page, so each old ad still costs one page read, and a strict filter on a list full of old ads may stop the run early to stay within what it has earned.
• A keyword is a role page of the site (enfermeiro, motorista, auxiliar de enfermagem), not a free-text search. A city needs its two-letter state.
• 15 ads a page. 1,000 jobs with job pages take about 40 minutes (2.5 s a job in our platform test). A list-only run reads one page per 15 ads and has no description and no date.
• employmentType is set only for contract words that name a type: "Efetivo/CLT" (a permanent job under Brazilian labour law) says nothing about hours, so it stays as text.
• Not given: apply link (the apply button needs a login), employer page, postal code, end date, experience, education. The site's total count is the national count of the role, not of your city, so it is not reported.
• Trabalha Brasil's robots.txt shuts out named AI-training crawlers. If you plan to train models on these rows, read the site's terms first.

Who uses it

• Recruiters and staffing agencies in Brazil: new ads by role and city every day.
• Pay research: the pay employers posted, in BRL with its period, without the site's role estimates mixed in.

Output: one row per job

Every Actor in this family writes the same columns, so you can merge boards in one sheet or database.

Field	What it holds
jobId	The board's own job id
title, company	Job title and hiring company (as the board shows them)
companyUrl, companyLogoUrl	Company page and logo on the board, when given
location, city, region, postalCode, country	Location as shown, and its parts; country is an ISO code (FR, DE, US...)
remote, workplaceType	true/false and remote, hybrid or onsite, only when the board says so
salaryMin, salaryMax, salaryCurrency, salaryPeriod, salaryText	Salary range as numbers, ISO currency, HOUR/DAY/WEEK/MONTH/YEAR; empty when the board gives none (never guessed)
employmentType, employmentTypeText	FULL_TIME, PART_TIME, TEMPORARY, INTERN, APPRENTICE, CONTRACTOR, FREELANCE..., and the board's own wording
experienceLevel, category, skills	When the board gives them
datePosted, validThrough	ISO 8601 dates
descriptionText, descriptionHtml	Full description as plain text and as HTML
applyUrl, jobUrl	Where to apply, and the job's page on the board
sourceBoard, searchLabel, scrapedAt	Which board, which of your searches found it, when
duplicateKey	The same key for the same ad posted on several boards (or reposted): title without (m/w/d) or H/F tags, company without legal form, city, country. Use it to merge results from several of our board Actors
extra	A few fields only this board has

Contact email addresses and phone numbers inside descriptions are removed: this Actor never collects recruiters' personal contact details.

Job alerts: only new jobs, sent where you work

1. Switch on Only new jobs since the last run and schedule the Actor (daily, hourly). It remembers the jobs it has already delivered for the same searches and filters, in a key-value store on your own account (named job-boards-seen-
2. Paste a Slack or Discord webhook (or a Zapier or Make webhook, which can email you) into Send new jobs to. After each run that found new jobs you get one message with the count and up to 10 jobs with their links. Free: the message is never billed.

Apify's own integrations (Google Sheets, email, webhooks) work too.

Pricing

Pay per job saved: USD 0.59 per 1,000 jobs (USD 0.00059 a job), plus USD 0.002 per run start. No monthly rental, no charge for jobs skipped as duplicates, already seen or filtered out. Set a maximum cost per run in Apify and the Actor stops just before it, keeping everything saved so far.

Use it through the API

curl -X POST "https://api.apify.com/v2/acts/usefulapi~trabalha-brasil-scraper/run-sync-get-dataset-items?token=YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords":["enfermeiro"],"city":"São Paulo","state":"sp","maxJobs":20}'

Python: ApifyClient("YOUR_API_TOKEN").actor("usefulapi/trabalha-brasil-scraper").call(run_input={...}), then read the dataset items.

Good to know

• Polite by design: one request at a time per site with a pause between requests, and the site's robots.txt is followed. Runs read public job pages only: no login, no candidate profiles or CVs.
• What a run summary tells you: the OUTPUT record has jobs saved per search, jobs skipped (duplicates, seen before, too old) and why a run stopped early, if it did.
• Is it legal? The Actor reads job ads that the board publishes openly, the same pages a job seeker sees. You are responsible for how you use the data (for example, the board's terms and data protection rules in your country).
• Found a problem? Open an issue on the Issues tab: we answer quickly.