Methodology

1. Where the listings come from

RápidoEmpleo does not post its own listings and does not accept paid job ads from employers. Every opening arrives through the official APIs of six partner boards, each with its own profile:

• Jobicy — remote jobs, mostly in English.
• Remotive — remote tech and product roles.
• Remote OK — global remote jobs.
• Jooble — on-site and hybrid roles by country.
• Adzuna — listings with structured salary data.
• Careerjet — broad coverage in Spain and Latin America.

That mix explains our catalogue's bias: coverage of remote and tech roles is better than of on-site trades, and coverage of Spain, Mexico and Brazil is better than of the rest of Latin America.

2. How often it refreshes

An automated job queries all six sources every 12 hours. Nothing is curated by hand: what you see is what the sources returned on the last pass.

3. What happens to a listing before it's published

The sources don't speak the same language: one says "Ciudad de México, MX", another "CDMX", another "Remote — LATAM". Before storing a listing we run it through our own normalisation:

• Country: free-text locations resolve to an ISO code, or to a region ("worldwide remote", "LATAM", "EMEA") when the role isn't tied to one country.
• Listing language: detected from the title and description, so you can filter by the language you'll have to apply in.
• Category and seniority: the role is classified into one of our 16 fields and, when the title allows it, into a level (junior, mid, senior, lead).
• Work mode: remote, hybrid or on-site, inferred from location, title and description.
• Text clean-up: unsafe HTML is stripped and the broken characters some sources send are repaired.
• Duplicates: the same opening often shows up in several sources. We compute a fingerprint from title, company and country and keep a single copy.

4. When a listing disappears

A listing that hasn't reappeared in any source for 30 days is marked expired and drops out of search. Its page stays reachable for a year — so a saved link doesn't dead-end — clearly marked as no longer active, and is then deleted.

Even so, a board can close a process without pulling the ad from its API. That's why every listing always links to the original posting: that's where you apply and where the real status lives.

5. How we calculate salaries

The figures in the salary guides and in the "Market snapshot" block are not estimates or survey data: they come from the salaries the listings themselves disclose. The calculation works like this:

• Listings without a salary are discarded. Today only 20% disclose one, so the sample is always smaller than the total number of openings.
• Everything is converted to an annual figure: monthly pay ×12, hourly rates ×2,080 (40 h × 52 weeks).
• Figures are grouped by currency and only the segment's most common currency is used, so pesos are never added to euros.
• We publish three numbers: the 25th percentile, the median and the 75th percentile. The median beats the average here, because a handful of very high salaries can't drag it.
• If a segment has fewer than 5 listings with salary data, we publish no figure at all rather than an unreliable one.

6. Limitations we know about

No aggregator has a complete picture of the market, and neither do we. These are the caveats worth keeping in mind when reading our data:

• The sample is the jobs advertised on our sources, not the whole labour market. A large share of hiring happens through direct contact and never reaches a board.
• Disclosed salaries skew toward roles and countries where pay transparency is the norm. With 20% of listings reporting pay, the medians are an indication, not a definitive benchmark.
• Category and seniority classification is automatic. It's reliable in aggregate, but any single listing can be mislabelled.
• The "city" field arrives messy from the sources (districts, regions or countries in the city slot). We group by canonical name, but we don't always get it right.

7. Errors and corrections

If you spot a duplicate listing, an opening that's already closed, a salary that looks wrong or a bad classification, write to us from the contact page with the link. We fix the data and, when the fault is ours rather than the source's, we adjust the rules so it doesn't happen again.