Start with one query and one location
The public starter searches recent AI engineer jobs in the United States and caps the output at ten rows. Edit the role and location only after checking that the first Dataset contains the fields your workflow needs.
{
"queries": [
"AI engineer"
],
"locations": [
"United States"
],
"includeDescriptions": true,
"maxResultsPerQuery": 10
}Inspect source identity before scaling
Check jobId, title, company, location, posting-date fields, and jobUrl. The Actor deduplicates by LinkedIn job ID across every query and location in the run, so one listing matched twice is returned and billed once.
RUN_SUMMARY diagnostic are not billed.Choose search, details, or specialist evidence
| Your starting point | Use this workflow |
|---|---|
| A role and location | Use Jobs Search for broad public discovery and optional description enrichment. |
| Known LinkedIn job URLs or IDs | Use LinkedIn Job Details Scraper for one complete record per supplied job. |
| Contract, freelance, temporary, or fixed-term work | Use LinkedIn Contract Jobs Scraper for direct classification evidence. |
| Salary, remote, sponsorship, skills, or company-level demand | Choose an evidence-backed intelligence workflow from the LinkedIn jobs collection. |
Export or automate after one clean run
Save a validated input as an Apify Task, schedule a daily or weekly snapshot, send results to a webhook, or retrieve the Dataset through the API. Snapshot runs bill the current matching rows again; use a dedicated monitor when you need a free baseline and only supported new-job events afterward.
Hosted MCP endpoint: https://mcp.apify.com/?tools=neuton/linkedin-jobs-search-scraper
Read the API guide for token handling, asynchronous runs, polling, and validated Dataset retrieval.
Frequently asked questions
Does it require a LinkedIn account?
No. It processes public job-search and job-detail surfaces without LinkedIn login credentials or cookies.
Can I paste an existing LinkedIn Jobs search URL?
Yes. The Actor accepts public LinkedIn Jobs search URLs, keeps only allowlisted search filters, discards tracking parameters, and never fetches an arbitrary supplied host.
Why might a successful search return fewer rows?
Duplicate job IDs are removed, and a narrow query, location, or recent-date filter can legitimately expose fewer public listings.
What should I run first?
Use the ten-result remote AI jobs Task, inspect the Dataset and free RUN_SUMMARY, then change one input dimension at a time.
Test ten public job rows
Use the bounded Task, verify every source field, and increase the limit only after the output fits your workflow.