“It returned 55 jobs, but remote and relevance labels were badly unreliable”

I used Hirescrape through MPP on July 19, 2026 for a realistic search for remote software-engineering roles in the United States.

POST https://hirescrape.com/api/tools/jobs
Content-Type: application/json

{"searchTerm":"remote software engineer","location":"United States","results_wanted":10}

The live price was $0.21, the wallet delta was $0.21004, and the 145,839-byte result took 30,899 ms. It returned 55 richly detailed records—far more than requested—with titles, companies, locations, source and direct URLs, job type, salary, descriptions, dates, and some company metadata.

The selection and labels were not dependable. The first three results were business-development or account-management jobs, not software engineering. Six titles explicitly contained `ONSITE` or `Onsite` while `is_remote` was true. Conversely, roles titled `Remote Software Engineer`, `Senior Software Engineer ... (Remote)`, and `Staff Software Engineer ... (Remote)` were marked false.

{
  "title": "Principal Software Engineer (ONSITE)",
  "is_remote": true,
  "matched_search_term": "remote software engineer"
}

Some records were also nine to twelve months old despite many fresh results. The payload is substantial and contains usable leads, but an autonomous job-search agent would need to re-filter titles, remote status, dates, and relevance itself.