Remote OK Jobs Scraper

Scrapes remoteok.com's official public JSON feed over plain HTTP — no browser, no login, no cookies. This is an invited integration: Remote OK publishes and documents the feed itself.

Read this first: a data-quality warning

Remote OK's free public feed does not match its website. In a live check of the first 12 listings, none were actually remote roles — they were on-site jobs like Store Manager, Bunbury, Removalist Offsider, Albury NSW, Orderly, Alice Springs and Loss Prevention Officer, Malé. Every one of them still arrives branded as remote, with a slug prefixed remote-.

This Actor does not pass that branding through as fact. It reads each listing's own title, tags and description and reports:

• location.remote — the evidence-based judgement
• board_claims_remote — always true, the board's assertion

Both are emitted so you can see the disagreement and decide which to trust. The verifiedRemote dataset view filters to rows where the text supports the claim.

The board's tags are similarly unreliable: digital nomad is auto-applied to listings including on-site retail jobs, so the detector ignores it.

If you need genuinely remote roles at volume, use the Himalayas Remote Jobs Scraper instead — same suite, same schema, ~100,000 listings whose remote status holds up.

Attribution is a licence condition, not a courtesy

Remote OK's API terms, returned in the feed itself:

Please link back (with follow, and without nofollow!) to the URL on Remote OK and mention Remote OK as a source, so we get traffic back from your site. If you do not we'll have to suspend API access. Please don't use the Remote OK logo without written permission as it's a registered trademark, please DO use our name Remote OK though.

Two obligations follow, and this Actor is built around them:

1. Every row carries apply_url (the canonical Remote OK link) and an attribution block, so the obligation travels with the data. If you republish, render that link as a followed link and credit Remote OK.
2. The Remote OK logo is never emitted. company_logo is the advertiser's own asset, which is a different thing.

What this Actor adds

Repaired text. The feed returns double-encoded text — Bachelorâs where the advertiser wrote Bachelor's. Titles, companies and descriptions are repaired, and the repair is applied only when the tell-tale byte sequences are present, so genuinely accented text is not corrupted.

Readable descriptions. HTML fragments are flattened to plain text with paragraph breaks preserved; the original HTML is kept alongside.

Honest salary labelling. is_estimated is true whenever a figure is present, because Remote OK derives these bands itself rather than taking a number the advertiser typed.

One schema across boards. Same record shape as the SEEK, JobStreet and Himalayas Actors, with a source-prefixed job_id.

Input

Field	Type	Notes
queries	array	Keywords over title, company, tags, description. Empty = whole feed.
tags	array	Weak filter — see the tag warning above.
maxItems	integer	The feed holds ~100 listings total; asking for more returns all of them.
includeDescription	boolean	Off gives a smaller dataset; the remote check is unaffected.
incremental	boolean	Only listings not seen in previous runs.
proxyConfiguration	object	Datacenter is enough.

Known limits

• No pagination and no query parameters. The public feed is a single snapshot of roughly the latest 100 listings. Keyword and tag filters are applied to that snapshot client-side. maxItems can only narrow it, never extend it.
• Feed contents are largely not remote. See the warning above.
• No structured location. The feed gives a bare place string, so country_code and region are null.

Sister Actors

Actor	Covers
Himalayas Remote Jobs Scraper	100,000+ worldwide remote roles — the better choice for remote work
SEEK Jobs Scraper	Australia, New Zealand
JobStreet & JobsDB Jobs Scraper	Indonesia, Malaysia, Singapore, Philippines, Thailand, Hong Kong

Local development

pip install -r requirements.txt
python test_local.py --max 20 --out sample_output.json
python test_local.py engineer --tags python --max 10

sample_output.json in this folder is real output from a live run, kept so the schema — and the data-quality problem — can be reviewed without running anything.