Remote Jobs from 4 Public Job Feeds in One Table

A remote jobs scraper and remote jobs API in one Actor: pulls Remote OK jobs and We Work Remotely jobs, plus Jobicy and Arbeitnow, from each site's own public feed, and returns one row per listing — deduplicated across sources — instead of you writing four separate integrations.

Introduction

Four public job feeds, read live on every run, each with its own shape: Remote OK's JSON API, Jobicy's JSON API, Arbeitnow's paged JSON API (kept to its remote: true rows only — most of that board is on-site), and We Work Remotely's public RSS feed. This Actor normalizes all four into one row shape, drops exact company+title duplicates across sources (keeping the earliest-posted copy), and pushes nothing invented: a source that fails or is blocked contributes one honest error row instead of guessed data.

A fifth source, Remotive, is listed in the input because it is a named source for this tool, but it is never actually fetched: Remotive's own robots.txt disallows /api/*, which covers the job feed endpoint. This Actor does not call a robots-disallowed path, so selecting Remotive always returns one row explaining that instead of live data. See SOURCE_TERMS.md.

Input

Field	What it does
sources	Which feeds to read. Default: Remote OK, Jobicy, Arbeitnow, We Work Remotely.
keywords	Optional. Keep only jobs whose title or tags contain at least one of these words (case-insensitive).
postedWithinDays	Optional. Drop jobs posted more than this many days ago.
maxItems	Hard ceiling on rows returned (1–5000, default 200). This is also your spend cap.
previousIds	Optional. dedupe_key values from an earlier run; matching jobs come back with is_new: false.
proxyConfiguration	When set, every request goes through that proxy. Off by default.

Pricing

$0.002 per row returned. No start fee. A run capped at 200 rows costs at most $0.40; a quiet run that returns fewer rows costs less. maxItems is the spend cap: the run cannot cost more than that number times $0.002, whatever the feeds hold.

Every row the run returns bills at that price, and that includes an error row (at most one per source that failed or, for Remotive, is robots-blocked). Error rows are counted against maxItems along with job rows, so the cap is a cap on the whole run, not just on the job listings. If the cap is small enough that both do not fit, room is kept for the error rows first, because they are what tells you a source was down.

Output — one row per job

Field	What it holds
source_name	"Remote OK", "Jobicy", "Arbeitnow", "We Work Remotely", or "Remotive"
source_url	The exact job page on the source's own site, unmodified
apply_url	The source's own URL for this job — never rewritten by this Actor. (Jobicy's free feed itself returns its own listing page here, not the employer's ATS link; see SOURCE_TERMS.md.)
job_id	The source's own ID or slug for this job
title	Job title
company	Employer name
company_logo	Always null — this Actor never ships a source's logo image (Remote OK's terms specifically reserve its logo)
location_limits	Where the source says this role is open to, as published (may be a country, region, or "Anywhere")
job_type	e.g. "Full-Time", "Contract", when the source publishes one
seniority	e.g. "Senior", "Director" — only Jobicy publishes this; null from the other three
salary_min / salary_max / salary_currency	When the source publishes a number; a 0/missing value is treated as unpublished, not a real $0 salary
tags	The source's own category/skill tags, as a list
posted_at	ISO 8601 UTC timestamp
description_text	Plain text (HTML tags stripped), capped at 5000 characters
dedupe_key	Normalized company|title, used to drop exact duplicates across sources and to match previousIds
is_new	true unless dedupe_key was in previousIds
fetched_at	When this run read the source, in UTC
error	null on a real job row; on a source-failure row, every other field is null and this holds the reason

Example (real row, live run 2026-09-21, Jobicy)

{
  "source_name": "Jobicy",
  "source_url": "https://jobicy.com/jobs/153812-creative-motion-design",
  "apply_url": "https://jobicy.com/jobs/153812-creative-motion-design",
  "job_id": "153812",
  "title": "Creative (Motion Design)",
  "company": "Superside",
  "company_logo": null,
  "location_limits": "Anywhere",
  "job_type": "Contract",
  "seniority": "Senior",
  "salary_min": null,
  "salary_max": null,
  "salary_currency": null,
  "tags": ["Creative & Design"],
  "posted_at": "2026-09-21T13:59:59Z",
  "description_text": "Superside is looking for a motion designer with the creativity, technical expertise, and innovation mindset to push the boundaries of motion design. [...truncated for this README; full run returned 3005 characters here]",
  "dedupe_key": "superside|creative motion design",
  "is_new": true,
  "fetched_at": "2026-09-21T17:23:58Z",
  "error": null
}

Required credit (please keep this)

Remote OK's and Jobicy's own terms require a kept, followed link back to the source on any page that shows this data, and Arbeitnow's and We Work Remotely's terms ask for the same. Every row already carries source_name and an unmodified source_url/apply_url for exactly this reason — if you build a page or app from this data, keep that link visible and pointing at the source's own URL, and do not swap in a different logo for Remote OK's. See SOURCE_TERMS.md for the exact quoted terms.

Recommendations

Run with keywords set to your role/skill so a small maxItems still returns a relevant table. If you only want Arbeitnow's genuinely-remote listings, expect a slower per-row yield from that source specifically — see "Limitations."

Limitations — what this does not do

• Remotive is never fetched. Its robots.txt blocks the feed path; see above and SOURCE_TERMS.md.
• Arbeitnow's remote rows are rare and can trip rate limiting. A live page of 250 Arbeitnow listings held only 12 marked remote: true on 2026-09-21, so this Actor must scan many pages per remote row. A live test run on 2026-09-21 was answered with an HTTP 429 Cloudflare challenge after 17 Arbeitnow page requests in one run; fetch.py now stops itself at 12 Arbeitnow page requests per run to stay clear of that. A run that asks for many rows and selects Arbeitnow may still get fewer Arbeitnow rows than other sources, or an Arbeitnow error row, rather than a guess.
• Jobicy's apply_url is Jobicy's own listing page, not the employer's application form, because reaching the employer's own link requires a paid Jobicy API key this Actor does not use. See SOURCE_TERMS.md.
• Jobicy can contribute at most 50 rows to any run. This Actor reads Jobicy's free feed in one call at its own page size (count=50); it does not page Jobicy for more, so raising maxItems never gets you more than 50 Jobicy rows in a single run.
• No claim about uptime or how fresh a feed is beyond what was measured in LOCAL_RUN.md at the time this Actor was built. Feeds update on each source's own schedule, not this Actor's.
• No de-duplication against your own database — previousIds only flags is_new; it does not filter rows out.
• When the proxy setting is set, every request goes through that proxy.
• No salary normalization. salary_min/salary_max/salary_currency are passed through as each source published them (period — hourly vs. yearly — is not separately captured); most Arbeitnow and We Work Remotely rows carry no salary at all.

FAQ

Where does the data come from? Four sources' own public feeds, read live on every run: Remote OK's JSON API, Jobicy's JSON API, Arbeitnow's paged JSON API, and We Work Remotely's public RSS feed. Never from a stored copy.

Does it need an API key or a login? No. All four feeds used here are public and unauthenticated.

Why did I get a row with error set and every other field blank? That source failed to answer, or (for Remotive) is permanently robots-blocked. No job was invented to replace it.

Why did I get fewer rows than maxItems? Either the selected sources and filters held that many jobs and no more, the run hit maxItems first, (for Arbeitnow specifically) the 12-page politeness ceiling was reached before enough remote: true rows turned up, (for Jobicy specifically) its own feed caps at 50 rows per run, or a small number of listings were dropped because the source gave no usable source/apply link for them (see "Required credit" above).

Is there support? Yes. Raise it through the Apify Store issues tab for this Actor, or email operations@ustechautomations.com. Replies within 2 business days.

Refunds and support

Refunds: if a run returns zero job rows for a valid input, email operations@ustechautomations.com within 7 days and we refund that run. Support: same address, replies within 2 business days.