CareerJunction Jobs Scraper — South African Job Postings

Collect job postings from careerjunction.co.za, a South African job board. Paste search results URLs with your own keywords and filters, category or location pages, or single job URLs, and get one record per job: title, company, location, salary as displayed with ZAR amounts, contract type, seniority and posting date.

Use dated exports to track a market, match postings by jobId, or monitor a sector over time. This Actor reads the public result pages.

Try it with 20 jobs

1. Paste a search results URL into Start URLs.
2. Set Maximum number of jobs to 20.
3. Start the Actor and open its dataset.
4. Read Run summary before exporting JSON, CSV or Excel: a successful run can still contain incomplete URLs or a listing the site cut short.

{
  "startUrls": [{"url": "https://www.careerjunction.co.za/jobs/results?keywords=engineer"}],
  "maxItems": 20
}

The form's job-count prefill is 20; the schema default is 0 (no limit). A broad keyword like engineer can span many pages; the number of jobs changes over time — set maxItems explicitly in API calls to bound the run.

Choose your start URLs

Input	Data	Notes
Search /jobs/results?keywords=…	Jobs matching the search, with the filters exactly as you chose them on the site (location, salary, contract type...)	Keep the URL the site gave you; the Actor collects from page 1 and follows pagination to the last page announced.
Category or location /jobs/<name> (e.g. /jobs/it, /jobs/cape-town)	Jobs of that category or city	Same listing format, same pagination.
Single job /<slug>-job-<id>.aspx	One job record, built from the job page's structured data	A job collected this way wins the deduplication against its listing-card duplicate.

You can mix all three. Jobs are deduplicated within the run by jobId. An invalid or disallowed URL is reported and skipped; it never blocks the others.

Full details (includeDetails)

For listing URLs, the Actor reads result pages by default, usually 25 jobs per page; fields absent from the page remain null. Enable Open each job page for full details to additionally open each delivered job's page once, for:

• description: the complete job description text,
• exact salaryMin / salaryMax from the site's structured data (it can be more precise than the displayed text),
• category: the published sector, e.g. agriculture, fishing & forestry,
• expiresAt: the exact expiry datetime, and the exact postedAt datetime.

This makes the run slower — one extra request per delivered job — and the Actor only does it for jobs it actually delivers (within maxItems). A job page that cannot be opened does not stop the run: that job is delivered without description and the run summary says how many pages failed.

Output

Field	Meaning
jobId	CareerJunction job identifier (deduplication key).
url, title	Job posting URL and displayed title.
company, companyUrl, companyLogo	Published employer or recruiting company, its profile URL and logo (logo links are signed by the site and can expire). Company names only — no personal contact details are extracted.
location	Location as displayed, e.g. Johannesburg.
salary, salaryMin, salaryMax, salaryPeriod	Salary exactly as displayed (R35,000 - R41,700 per month, R Undisclosed), numeric bounds in ZAR when the text writes amounts, and the pay period. Undisclosed salaries keep their text with null numbers. Some recruiters mask a salary as R01 - R02 per month: the numbers are delivered as displayed, the text tells you what the site actually showed.
currency	Set to ZAR by this Actor, which is designed for salaries displayed in South African rand; no currency conversion is performed.
postedAt	Posting date, ISO (2026-10-20 from the card; exact datetime with includeDetails).
jobType, seniority	From the displayed position line: Permanent / Contract / Temporary, and Junior / Senior / Specialist...
category, description, expiresAt	Job-page details; only filled with includeDetails or for a single job URL. Description text is delivered as published, including any contact text or masks supplied by the site; this is not an anonymisation service.
expiresIn	Expiry as displayed on the card (Expires in 5 days).
page, position, sourceUrl, scrapedAt	Provenance: listing page and card position, start URL, UTC read time.

An absent field is null, never an empty string.

Real output example

This exact listing record comes from run zJZp8JEJeEOtbXqaH (build 0.1.6), completed on 2026-10-22T22:12:15.656Z. It demonstrates the format, not today's postings. Signed logo URLs can expire.

{
  "jobId": "2646461",
  "url": "https://www.careerjunction.co.za/agriculture-engineer-uae-job-2646461.aspx",
  "title": "Agriculture Engineer – UAE",
  "company": "Hire Resolve",
  "companyUrl": "https://www.careerjunction.co.za/companies/24781/hire-resolve",
  "companyLogo": "https://careerjunction-web-prd.s3.eu-west-1.amazonaws.com/assets/employer-logos/24781-240x120.png?X-Amz-Expires=86400&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIA3UBW7IFPPL4T7PMH/20261022/eu-west-1/s3/aws4_request&X-Amz-Date=20261022T221204Z&X-Amz-SignedHeaders=host&X-Amz-Signature=5a59508e91f5ae6cdfdafc47c1a2b35ca44dfea7bcd59a9a4a632f5e83d73ad0",
  "location": "United Arab Emirates",
  "salary": "R75,000 - R91,700 per month",
  "salaryMin": 75000,
  "salaryMax": 91700,
  "salaryPeriod": "per month",
  "currency": "ZAR",
  "postedAt": "2026-10-21",
  "jobType": "Permanent",
  "seniority": "Senior",
  "category": null,
  "description": null,
  "expiresIn": "Expires in 33 days",
  "expiresAt": null,
  "page": 1,
  "position": 17,
  "sourceUrl": "https://www.careerjunction.co.za/jobs/results?keywords=engineer",
  "scrapedAt": "2026-10-22T22:12:05+00:00"
}

With includeDetails enabled, the same job carries "category": "agriculture, fishing & forestry", "expiresAt": "2026-11-24T07:12:22", the exact posting datetime and the full description text.

Run summary

The Actor leaves two readable records:

• OUTPUT (JSON): jobs delivered, per-URL statuses (done, no_jobs, stopped_at_limit, not_found, partial_collection...), rejected URLs with reasons, and automatic check warnings (anomalies, empty when everything is normal).
• Run summary (HTML): the same information as a table, in the console's Output tab.

A listing the site cut short — fewer jobs than the site announced — is reported as partial_collection with a check warning, never as a quiet success. A search with no results is no_jobs, not a failure. If no job is delivered and at least one accepted URL fails, is interrupted or cannot be parsed, the run fails with an explicit status message. A changed page structure is reported as structure_changed, or partial_collection if that listing already delivered jobs.

Troubleshooting and support

Check Run summary first. For partial_collection, compare the delivered and announced counts and inspect the affected start URL. For blocked_by_site, rate_limited or structure_changed, keep the run ID and the rejected/affected URLs when reporting the issue. A legitimate empty search is reported as no_jobs and has no result charge.

Report reproducible problems in the Issues tab. Include the run ID, input URLs and expected versus observed result; do not include API tokens or private credentials. Hexagone Data maintains this Actor. Repeated site changes or failures require investigation; a successful run with warnings does not establish full coverage.

Cost and operating limits

• Price: charged per delivered job ($1.50 per 1,000 results), never per request. For example, 20 delivered jobs cost $0.03 and 1,000 cost $1.50 in result charges; a run delivering no jobs has no result charge. Any additional platform or proxy costs depend on your Apify configuration. maxItems and your charge limit bound the run.
• Politeness: about one request every 2 seconds, starting with a plain HTTP client and escalating access if necessary. The Actor rejects /myprofile/ account URLs based on the robots.txt rules checked on 22 October 2026. Those rules are not refreshed during each run.
• Durability: if the platform migrates the run mid-collection, it resumes from its checkpoint without delivering duplicates.