SEEK jobs scraper: search seek.com.au and seek.co.nz by keyword and location and get current job ads with salary as min and max numbers. Typical run: a 50-job search in about 10 seconds. Built for AI agents and MCP: keywords, a location and a job limit in, flat JSON rows out in one call.
When seek.com.au / seek.co.nz refuse the address (HTTP 403), the same SEEK API continues on the sister host my.jobstreet.com; then no-proxy, then the unblocker. Free-plan refusals come as a free notice row instead of an error row.
2026-11-10
Salary parsing fixed on real adverts (275 Sydney nursing listings, 10.10–10.11.2026): "p.h.", "ph", "/hr", "p/h", "p/d", "p.a." and "/yr" now fill salaryPeriod (123 of 164 priced adverts instead of 100); cents are kept ($41.58 is no longer $42); a figure already written in thousands ("$133348.0k") is no longer multiplied again into $146,945,000 a year; a superannuation percentage ("+ 12% super") and a leave count ("6 Weeks AL") are no longer read as pay; salaryMin/salaryMax can no longer come back reversed.
2026-11-02
Fix: when "Max cost per run" stops a run before any row is delivered, the status message now quotes the price of the rows you asked for (not the cheapest row of the Actor), and never reports a row count over an empty dataset.
2026-10-21
Maintenance: run-end alerts updated to the current shared version.
2026-10-11
A dropped connection is retried on a fresh proxy address instead of failing the whole search.
"Max cost per run" is respected row by row, and the status line says when it stopped the run.