Remote Jobs Feed: daily new remote jobs, de-duplicated

Get a daily feed of new remote jobs from four public job feeds (Jobicy, Remote OK, Working Nomads, Arbeitnow) plus any company job boards you want to watch (Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Recruitee, Personio, Teamtailor, BambooHR, Workday, Breezy). Filter by keyword, country or region and employment type. The same job listed on several sources becomes one row. In new jobs mode each run returns only the jobs you haven't seen yet. Every row has the same 31 columns, so it drops straight into a spreadsheet, Airtable, a Slack alert or a job board.

Read this first: what it can and can't do

• It is only as wide as its sources. About 1,700 remote jobs a week (measured 1 Nov 2026, last 7 days, after merging duplicates) plus whatever company boards you add. That's not every remote job on the internet: LinkedIn, Indeed and others aren't included.
• Jobicy (most of the remote jobs) serves only the last 7 days, and its feed runs about 3 hours behind the site. Remote OK serves only its latest ~100 jobs and Working Nomads its latest ~35. Schedule the Actor daily and you won't miss them; a one-off run can't go back further.
• Arbeitnow is mostly Germany/Europe and mostly on-site. Its "remote" flag isn't reliable, so we also read "remote" / "Homeoffice" in the location and title.
• Excluded on purpose: Remotive, Himalayas and We Work Remotely forbid this kind of reuse in their terms.
• Location matching is text-based. "Australia" matches jobs whose location says Australia, APAC / Asia Pacific (unless you turn that off), or Worldwide / Anywhere. You can also enter Worldwide (only jobs open worldwide) or Remote (worldwide jobs plus jobs that just say "Remote"). Many Remote OK jobs give no location at all ("unspecified"). They are left out when you filter by location, unless you turn on includeUnspecifiedLocation. Always check the job page for the real country rules.
• The same job posted for several countries becomes one row. Its matching locations are joined, e.g. "Germany; France; Spain". The other postings' links are in otherJobUrls.
• Some company boards don't say if a job is remote (workplaceType unknown). Lever, Ashby and Workday boards don't publish the company name, so we use the board's name: "palantir" → "Palantir".
• "New" means new to your monitor. Feeds don't always list every job on every read, so a job can first reach you a day or two after it was posted. Compare firstSeenAt with postedAt.
• Working Nomads links go through Working Nomads to the employer's own job or apply page.
• Show it with credit. The free feeds ask that you name them and link to the job page. Every row has jobUrl, sourceName and an attribution line. Apply links for Jobicy, Working Nomads and Arbeitnow open the job page on that site.
• Bad input costs nothing. A wrong mode or source, a website instead of a board link, or a typo gets a plain-English message and a free run.

How to use it

1. Daily feed: set keywords (e.g. python, data engineer; matched in job titles at the start of words, so engineer also matches "Engineering Manager". Turn on searchTags or searchDescriptions to look wider) and locations (e.g. Australia). Leave mode on New jobs only and save the input as a task with a daily schedule. The first run returns the current matching jobs once (isNew = false). After that, each run returns only jobs that appeared since the previous run (isNew = true).
2. Everything right now: mode All matching jobs.
3. Watch companies too: add their board links to companyBoards, e.g. https://jobs.lever.co/acme or greenhouse:acme. Only have websites? Our Company Jobs Finder finds the board from a website.

The monitor's memory lives in a named key-value store (remote-jobs-<monitorName>). Jobs are remembered before filtering, so changing filters never makes old jobs look new. Jobs that a capped run (maxJobs or max cost) didn't return come back on the next run only if they are genuinely new.

Input example

{
  "mode": "new_jobs",
  "keywords": ["python", "data engineer"],
  "locations": ["Australia"],
  "excludeKeywords": ["senior"],
  "companyBoards": ["greenhouse:gitlab"],
  "monitorName": "python-remote-au"
}

Output (one row per job, always these columns)

Lists (tags, alsoListedOn, otherJobUrls, matchedKeywords) are plain text, so a CSV or Excel export always has exactly these 31 columns, in this order.

Column	Meaning
jobKey	Stable unique key (source + its job id)
title, company, location	As the source writes them
remoteScope	worldwide / restricted (limited to named places, see location) / unspecified (just "Remote") / null if not remote
workplaceType	remote / hybrid / onsite / unknown
employmentType	full_time, part_time, contract, internship, freelance, temporary or null
seniority, category, tags	When the source gives them
salaryMin, salaryMax, salaryCurrency, salaryPeriod, salaryText	When published
postedAt, ageDays	Earliest posting date across sources
jobUrl, applyUrl	Job page / application link
source, sourceName, sourceJobId	Where the row comes from
alsoListedOn, otherJobUrls	Other sources listing the same job (merged)
companyBoardUrl	For company-board jobs
matchedKeywords	Which of your keywords matched
isNew, firstSeenAt	New-jobs mode
descriptionText	Only with includeDescription
attribution, scrapedAt	Source credit line, run time

There are also two records in the key-value store. SOURCES shows every feed and board with its status, jobs read and any error. RUN_SUMMARY shows counts, jobs removed per filter, duplicates merged and what was charged.

Sample row:

{"title":"Senior Data Engineer","company":"WOO X","location":"APAC","remoteScope":"restricted","workplaceType":"remote","postedAt":"2026-10-28T14:53:27.000Z","jobUrl":"https://jobicy.com/jobs/157598-senior-data-engineer-6","sourceName":"Jobicy","matchedKeywords":"data engineer","isNew":false}

Pricing (pay per event)

Event	Price
Feed check: once per run in which at least one source was read	$0.0015
Company job board read (each board you add, per run)	$0.001
Job returned (all_jobs mode, or a current job on a monitor's first run)	$0.0015
New job returned (new_jobs mode)	$0.0025

Store discount tiers: free-plan and Bronze users pay the prices above. Silver gets 10% off and Gold (and higher) gets 20% off on returned jobs and new jobs (the per-run feed check and company board fees are the same on every tier).

Filtered-out jobs, merged duplicates, already-seen jobs, failed sources, invalid input and self-tests are free. Example: a daily feed that finds 5 new jobs a day costs about 30 × ($0.0015 + 5 × $0.0025) = $0.42 a month. Set a maximum cost per run in the run options for a hard cap.

Reliability

• Each source is read on its own. If one fails, the others still deliver; the failure is listed in SOURCES and is free. If every source fails, the run fails and nothing is charged.
• Self-test: run with {"selfTest": true} for a free health check: 34 checks (parsers on saved samples, location rules, filters, merging, monitor, input rules, plus a live read of every feed). We run it every day.

FAQ

Why did a first run return jobs marked isNew: false? That's the baseline: the jobs that existed before your monitor started. Turn off includeExistingOnFirstRun to start silent (the first run then charges only the feed check).

Can I get jobs older than 7 days? From Arbeitnow and company boards yes (set postedWithinDays, 0 = no limit). Jobicy, Remote OK and Working Nomads don't publish older jobs.

Can it apply for me or send e-mail alerts? No. Connect the dataset to Slack, e-mail or Google Sheets with Apify integrations, Make or Zapier.