Methodology · Updated 2026-09-11

How we aggregate remote jobs

RemNavi is an aggregator. We don't post jobs ourselves; we collect listings from 16major remote job boards and ATS platforms, deduplicate them, normalise the fields, and present them under one searchable interface. This page explains exactly how we do it, what our data means, and where it doesn't apply.

3,441 live · 10,300 tracked listings16 sourcesCC BY 4.0 data licence

01: Sources

Where our listings come from

Listings come from publicly accessible job boards, ATS-hosted career pages, and one labelled paid partner feed. We never scrape behind a login, and we always link through to the source so the application happens at the employer.

  • Greenhouse (40+ companies)

    ATS: direct career pages for Stripe, Anthropic, Datadog, Okta, Figma, Duolingo, and many more. Our single largest source by listing volume.

    greenhouse.io
  • 4 Day Week

    Board for compressed-week and work-life-balance-forward remote roles.

    4dayweek.io
  • Ashby (30+ companies)

    ATS: direct career pages for OpenAI, Snowflake, Cohere, Ramp, Notion, and many more.

    ashbyhq.com
  • We Work Remotely

    Long-running remote board; salary often free-text.

    weworkremotely.com
  • VueJobs

    Vue.js-focused remote job board.

    vuejobs.com
  • Jobicy

    Curated remote job board; transparent salary disclosure.

    jobicy.com
  • Remote OK

    High-volume remote board with structured salary fields.

    remoteok.com
  • Arbeitnow

    Remote and Europe-focused job board.

    arbeitnow.com
  • Working Nomads

    Curated remote and digital-nomad job board.

    workingnomads.com
  • Lever

    ATS: direct company career pages.

    lever.co
  • Remotive

    Remote-first board with strong SaaS employer coverage.

    remotive.com
  • Himalayas

    Remote job board with structured listing fields.

    himalayas.app
  • Python.org Jobs

    Official Python Software Foundation jobs board.

    python.org
  • Golang Projects

    Go-language-focused remote job board.

    golangprojects.com
  • Hacker News Who's Hiring

    Monthly thread, parsed for structured fields. Indexed when the current month's thread carries listings; between threads this source contributes zero.

    news.ycombinator.com
  • Adzuna

    Paid partner feed, not a free public board. Carries a visible Sponsored label on every listing, passes the same classification and scoring rubric as every other source, and never receives a ranking or score boost for being paid.

    adzuna.com

02: Cadence

How often listings refresh

Every source is polled on its own schedule: aggressive enough to catch new listings within hours, polite enough that we do not strain the origin.

Board-style sources with public APIs (Jobicy, Remote OK, Remotive) are ingested every 30 minutes. Greenhouse and Lever career pages are polled every 2 hours per employer - more frequent for the companies we cover as hubs. We Work Remotely and Hacker News threads are parsed daily.

A listing disappears from RemNavi when its source returns 404, when it has not been seen for 14 consecutive polling cycles, or when the employer marks it closed. The 14-cycle buffer guards against transient source outages producing false negatives.

03: Deduplication

How we handle the same job on multiple sources

A listing that appears on three boards is the same listing. Counting it three times is how aggregators inflate numbers. We don't.

Every listing is keyed by a stable hash of normalised_company + normalised_title + region. When the same hash appears on multiple sources, we keep the canonical ATS (Greenhouse or Lever) if present, otherwise the earliest-seen source, and we record the other sources as alternates on the listing record.

Normalisation is case-insensitive, strips legal suffixes (Inc., GmbH, Ltd.), collapses whitespace, and maps common title variants (e.g. "Sr." → "Senior", "PM" → "Product Manager") through a small managed dictionary. The dictionary is open in the repo and reviewed manually on every change.

Why you'll see two different totals on the site. The listing count on the homepage, /press/, /press-kit/, and /tools/salary-index/ is the deduplicated, browsable count: canonical postings only (duplicates from the same job on multiple boards excluded, per the dedup process above) plus, for work-mode, everything except listings explicitly classified onsite-only. This is what you'd actually see paging through /jobs/. The /trends/ "Listings tracked" chart intentionally shows a larger number: it's the full historical ingest total, including items not yet deduplicated or classification-pending, tracked day over day since recording began. We report it on its own definition rather than retrofitting the dedup-filtered count onto the historical series, since which duplicates existed on a given past day isn't something we recorded at the time, so recalculating it after the fact isn't possible without breaking the series.

04: Salary data

How the Remote Salary Index is built

The Salary Index surfaces only listings with transparent, structured compensation; we never fabricate ranges, infer them from proxies, or interpolate between employers.

A listing enters the Salary Index only if the source provides a numeric range or a single numeric figure. Free-text salary language ("competitive", "market rate", "DOE") is excluded. We convert every range to USD per year using a daily FX rate; we do not attempt to correct for cost of living at this layer (that's what the Salary Calculator is for).

We report median and range per role group and per seniority band. Groups with fewer than 5 listings in the current window are suppressed to avoid statistical noise. The published snapshot at /tools/salary-index is built at every release and exposes the row-level underlying data via CC BY 4.0.

Limitations we want you to know about. Our sample is biased toward employers who publish salary bands: in practice, that's US-headquartered SaaS, European remote-first companies, and European employers subject to the 2026 EU Pay Transparency Directive. US states with pay transparency laws (CA, CO, NY, WA) are overrepresented in the underlying data. Treat the medians as lower-bound signals of real compensation, not universal industry figures.

05: Boundaries

What RemNavi is not

Clarity about what we do not claim is more useful than hedge-everything disclaimers.

  • We are not a job board. RemNavi never sells ranking or score influence: paid or not, no listing's score or sort position can be bought. We do carry clearly labelled paid placement (Featured, Boost, and Spotlight slots) and a labelled partner feed (Adzuna), classified and scored the same as every other listing. All listings, free or paid, link through to the original employer page.
  • We are not a recruiting agency. We do not hold candidate profiles, we do not submit applications, we do not broker between candidates and employers.
  • Our salary figures are not a compensation benchmark. They are a live snapshot of what employers openly advertise, useful as a negotiation anchor, not as an authoritative industry rate.
  • Our "Real Remote Score" is an editorial rubric, not an endorsement. Employers cannot pay to influence their score.

06: Licence

Using our data

The structured data underlying the Remote Salary Index is published under Creative Commons Attribution 4.0. Use it, cite us, link back.

The Salary Index snapshot (CSV + JSON) is available at /tools/salary-index. A broader Market Index (source, role, skill, company, geo, and seniority cuts) is at /market-index, with direct JSON and CSV download links. Every listing is also scored on the Real Remote Score rubric. Attribution for any of these: "RemNavi, https://remnavi.com". Commercial reuse is permitted under the same CC BY 4.0 terms.

Formal citation: this Methodology is archived at DOI: 10.5281/zenodo.20234690. The State of Remote H1 2026 report is archived at DOI: 10.5281/zenodo.20322848.

Individual job listings remain the intellectual property of the source employer and posting platform. Our right to surface them is under fair use for aggregators; we do not store full job descriptions, only the metadata required for discovery and the canonical link back to the source. If you are an employer and would like a listing removed or a source excluded, write to hello@remnavi.com and we will action within 48 hours.

Transparency

Found something we got wrong?

We care about getting this right. If you see a duplicated listing, an outdated salary, or anything that contradicts the rules on this page, let us know and we will fix it.

hello@remnavi.com