Indeed Scraper — Indeed Jobs API with Salaries
Pricing
from $2.00 / 1,000 jobs
Scrape Indeed jobs in 22 countries by keyword, location or Indeed URL: title, company, rating, size, salary parsed to min and max, job type, remote flag, posting date, full description and apply link. No login, proxy included. Export CSV, JSON or Excel.
All notable changes to Indeed Jobs Scraper are recorded here. The newest release is first.
excludeAgencies now also catches company names with Talent, Placement, Employment Services, Workforce Solutions or a trailing HR (for example "Talent Spark HR"), and descriptions saying "staffing / recruiting agency". Names such as "HR Block" or "Workforce Software" are kept. Default off; no other output changes.excludeAgencies filter (default off, so existing inputs behave as before): skips jobs posted by staffing, temp and recruitment agencies, detected from the company's industry, its name (agency words and large agency brands) and the description's own wording ("Our client is...", "Personaldienstleister"...); phrases after a negation do not count. Skipped jobs are free; the run's status message says how many were skipped.New filters, sorting and monitoring. Existing inputs keep working; new output columns are added, none removed.
experienceLevel filter (entry, mid, senior), sent to Indeed as its own attribute filter, and a new experienceLevel column on every job Indeed tags with a level. Indeed URLs with sc=0kf:explvl(..._LEVEL); work too.salaryMin filter: minimum yearly pay in the site's currency (hourly, daily, weekly and monthly pay converted to a year; Indeed's estimate counts; jobs without salary skipped). New salaryYearlyMin / salaryYearlyMax columns.sort: relevance (default, unchanged) or date. Indeed's API ignores a date sort, so the Actor reads a pool of 3 x each search's share of maxItems (100-1,000 jobs) and returns the newest first. sort=date in an Indeed URL now does the same instead of being ignored.dedupeAcrossSearches (default on): a job another keyword or URL of the same run already returned is skipped, free, instead of saved and billed twice.onlyNewSinceLastRun + seenJobsStoreName: scheduled runs return and bill only jobs no earlier run delivered (memory in a named key-value store on the user's account).not_found row instead of failing.[name removed].First release.
not_found when it is gone); malformed keys and non-search pages give a free error row instead of an unrelated search.description and descriptionHtml before a row is saved. A long ID number in the text can also be removed; a phone number is never kept.status, so a search without matches (not_found) and a failure (error) are told apart from real data. Only ok rows are charged.