Editorial and data methodology
Useful evidence needs visible limits.
How we separate dated editorial judgements from a changing sample of vacancies.
Two different kinds of evidence
Board reviews interpret public product information. The live jobs pages and Remote Work Index summarise records returned by the Jobicy feed. Neither is a complete labour-market study, and we do not use the size of the current feed to calculate a job board’s editorial score. The scoring methodology contains the criteria, weights, rating anchors and current component marks.
For product facts we prefer official search pages, FAQs, pricing, terms and API documentation. We record the date checked, distinguish provider claims from our interpretation and identify unverified matters. Public source review does not establish the experience of a paid account or demonstrate hiring outcomes.
Where the current listings come from
WRILS requests up to 100 recent records per query from Jobicy. The site refreshes server-rendered samples no more frequently than hourly; interactive searches request the selected filters separately. We retain the source listing URL and show the provider attribution. This is a deliberate display sample, not a claim that the upstream API contains only 100 jobs.
The current upstream documentation supports cursor pagination and batch status checks within a recent-publication window. An earlier WRILS description incorrectly said these features did not exist. We corrected that description on October 17, 2026. WRILS does not currently traverse every cursor or retain a complete historical dataset. A vacancy can close between retrieval and a click, so the employer’s application page remains the final availability check.
What the index counts
The denominator is the number of jobs returned for the displayed sample. Company counts use the company names in those records; they are not verified counts of distinct legal entities. Region counts summarise source labels, which can overlap or describe different geographic levels. A category count can include a listing in more than one category. These limitations mean that a chart is useful for understanding the sample’s composition, not for ranking national labour markets.
Salary disclosure means that at least one salary-bound field is present in the record used by the page. It does not mean the range is narrow, the currency is comparable or the pay applies to every location. We do not replace a missing salary with a guessed zero or treat an advertised maximum as an offer. See our salary-transparency guide for a more careful reading procedure.
Freshness and failed requests
Feed timestamps describe the returned data, while a page’s last-reviewed date describes editorial work. They answer different questions. If a request fails, a page can show an empty state; zero displayed records should not be read as proof that there are no remote jobs. Repeated imports, inconsistent titles and source publication dates can also affect the apparent distribution.
We do not draw historical trend lines from a single rolling sample. A defensible trend would need scheduled snapshots, a stable collection method, ID-based deduplication and documented treatment of closed or missing records. The current index has not established that historical series. Its chart labels and explanations should be read as a snapshot.
Examples, advice and editorial judgement
Illustrative workflows and candidate examples are written to explain a decision, not to imply that a named person obtained a particular result. A subscription recommendation is conditional on role fit, eligibility and total commitment. A board’s screening claim does not replace checking the employer. No score predicts acceptance into a talent marketplace, an interview invitation or a safe outcome.
Where product access or pricing is unclear, we point to the official page instead of inventing a definitive limit. A correction can change our description without implying that the provider changed its product. Scores rebased under a new methodology should not be interpreted as directly comparable measures of product performance over time.
Corrections and source attribution
We identify the source of live listing data and cite the official product pages used in reviews. Send a page URL and supporting evidence through Contact if a claim is inaccurate. Material scoring changes are documented on the affected review; privacy practices are described separately in the privacy policy.
Official upstream reference: Jobicy API, RSS and MCP documentation. This reference describes available upstream capabilities; it does not imply that WRILS uses every capability.