We Work Remotely Jobs Scraper - Remote Jobs by Category

Remote job postings from We Work Remotely's public RSS feeds, as flat, typed dataset rows: job title, company, category, job type, region, allowed countries, posted and expiry dates, company website and headquarters, and the full description as both HTML and plain text. Pick categories, filter by keyword, run it on a schedule and send the dataset to Sheets, a webhook or your warehouse.

What you are paying for, honestly

The data comes from the RSS feeds We Work Remotely publishes itself. They need no login, cookie or API key; anyone can read the same XML. What this actor adds:

• Flat, typed rows. The company is split from the title (Vercel: Site Engineer becomes company and title), dates are ISO 8601, the Headquarters and website lines are lifted out of the description into their own columns, the country list is an array without flags, and the description is also given as clean text.
• Several feeds in one run, deduplicated. The same posting found in two feeds, or listed twice in one feed, is one row and one charge.
• Keyword filter over title, company, category, skills and description, applied before the row cap, so jobs you filter out are not billed.
• Refusal handling. A refused (HTTP 403/429/5xx), timed-out or non-feed response is retried from a fresh proxy address; an unknown category is reported and skipped.

Input

Field	Type	Default	What it does
categories	array	remote-design-jobs	Category feed slugs, one request each. Empty = the all-jobs feed. A full feed URL also works.
keywords	array	-	Keep jobs mentioning any of these terms (case-insensitive).
maxItems	integer	100	Most jobs for the whole run, after dedupe and keywords. Each job is billed, so this is a spending cap.
proxyConfiguration	proxy	Apify proxy	We Work Remotely answered 200 from an Apify datacenter and a residential address on 2026-10-09; the proxy is there so a refused request can be retried from another address.

Which feed is which (measured 2026-10-09):

• The all-jobs feed (leave categories empty) is a sampler: 89 jobs, about the newest 10 per category.
• Category feeds are deeper: remote-sales-and-marketing-jobs 89, all-other-remote-jobs 72, remote-customer-support-jobs 52, remote-full-stack-programming-jobs 50, remote-design-jobs 29, most others 18 to 25, remote-front-end-programming-jobs 5. Feeds hold the board's current postings, not its history.
• Slugs, as on the site: remote-customer-support-jobs, remote-design-jobs, remote-product-jobs, remote-full-stack-programming-jobs, remote-front-end-programming-jobs, remote-back-end-programming-jobs, remote-devops-sysadmin-jobs, remote-sales-and-marketing-jobs, remote-management-and-finance-jobs, and all-other-remote-jobs (no remote- prefix on that one). Slugs are lower-cased for you. A slug We Work Remotely does not have is answered with an empty HTTP 301; the run logs it, skips it and carries on.

Output

One row per job posting.

job_id, guid, url                  job_id = last part of the URL; guid = the feed's guid (the URL)
title, company                     "Vercel: Site Engineer" -> company "Vercel", title "Site Engineer"
category, job_type                 "Design", "Full-Time" / "Contract"
region, countries, state           where the role is open (see below)
skills                             the employer's tags, as one string
published_at, expires_at           ISO 8601 with offset, e.g. 2026-10-08T07:31:34+00:00
headquarters, company_url, company_logo_url
description_html, description_text
source_feed                        "all" or the category slug that produced the row

Dataset views: Jobs and Jobs with descriptions.

Null means "not given", never a guess

Counts are over the 356 distinct postings read from 13 feeds on 2026-10-09.

• company splits at the first ": " of the feed title, so IxDF - Interaction Design Foundation: Course Writer and Editor: UX, UI, and AI has company IxDF - Interaction Design Foundation. A title without that separator would give company null and the whole text as title; none of the 356 did. A company name that itself contains ": " would be cut short.
• state (251 of 356 null) and skills (266 null) are empty on most postings. That is normal. skills is left as one string, because the tags themselves contain "and" (Email Marketing and Automation) and cannot be split reliably.
• countries is an empty list, not null, when the employer named none (269 of 356). When set it is the countries the employer will hire in, flags removed, as WWR writes them (United States of America, Korea (Republic of)). It narrows region, which is usually Anywhere in the World and sometimes a city or EMEA.
• headquarters (7 null) and company_url (205 null) are read from the first paragraph of the description, where We Work Remotely prints them; null when the employer left them out. company_logo_url (258 null) is sent on about one posting in four.
• job_type is null when the feed sends an empty <type/> or no type tag, which happened for 1 of the 356 and for every posting of the legacy remote-programming-jobs feed. published_at and expires_at are null only if the feed omits the tag; the legacy feed omits expires_at.
• description_text omits the Headquarters/URL paragraph and the logo (they have their own columns). description_html is as sent, except that an inline data: image is emptied (src=""): one real posting embedded a 1.4 MB base64 PNG.
• We Work Remotely lists a posting once per location variant under one URL (measured: Tailscale's Backend Engineer, Control Plane with Headquarters

  Remote (United States)
  and again with Remote (Canada), same guid, same feed). Rows are deduplicated by guid and the first copy wins, so the other variant is not returned.

Pricing

Pay per event: a small actor start fee plus one job scraped event per row. Set maxItems to cap spend.

Limits

• Only what the feeds list: the current board, not archived postings, and not every posting on the site (the all-jobs feed in particular is a sample).
• A snapshot: postings expire (expires_at) and are added through the day.
• Not affiliated with We Work Remotely.