A Spanish company-lookup site that's been running for 11 years: free for nine years, and only after adding a signup wall did they find their users were thousands of businesses

LibreBOR split the PDFs of Spain's daily business registry bulletin into millions of "company pages" and "executive pages", living off search traffic for ten years before it started charging in 2024. Its global ranking recently rose 2.6x, but most of that climb probably isn't real people. Two things are worth watching: how it built that traffic, and how it turned traffic into money. Also note that by 2026 cheaper APIs were already going after its customers.

A 7-day legal demand letter

On 2019 Mar 12, Spanish programmer Pablo Castellano received a registered letter from a lawyer. It came from the trademark agent for AEBOE, Spain's official gazette office, demanding that he stop using the name "LibreBORME" within 7 days, since "BORME" is a trademark the gazette office holds (they've registered the word across nearly every class, from chemicals and cosmetics to tobacco), and that he hand over the libreborme.net domain for free.

The site had been live for 4 years by then. It began as his university thesis project: Spain's Commercial Registry Bulletin (BORME) publishes new company formations, director changes, capital increases and dissolutions every day, but the official source only offered PDFs sorted by date, with no search and no way to look up "which companies has this person been a director of". He wrote a parser that split each PDF apart and built a page for every company and every executive. The code is open source under AGPL.

He didn't end up in court. He renamed the project LibreBOR and kept the site running. He wrote up the episode on his blog, and it circulated again among transparency advocates and journalists.

What it actually is

A Spanish company and executive search site, plus an API. Enter a company name, tax ID (NIF) or person's name to see the company's registry history over the years, current and former directors, registered address and industry codes. Searching by person shows every company where they've held a role. To get a company's full history officially, you have to pay the provincial commercial registry for a request, usually 20–30 euros, and wait 3–5 days (per the mapasocietario.es comparison page).

According to the official blog from January 2025, the database holds over 300 million companies and 3.3M executive positions, and every record is its own page that Google can index, with an English version too.

Where users come from: millions of pages, plus ten years of word of mouth

• Long-tail search, the main channel. People checking a partner, supplier or former boss search "company name + administrador" or "company name + NIF", and often land on its pages. In Chrome real-user data, it sat in the "top 50k sites" tier in Spain every month from May through September 2026.
• Early press and community. El Confidencial wrote an article in February 2016 titled "Free tools for investigating Spanish companies from home"; the Catalan Wikipedia has an entry for it; and Spain's government open data portal datos.gob.es listed it in its app catalog.
• A steady stream of tutorials. The OSINT investigation site ciberpatrulla.com runs a LibreBOR usage guide (written by 2025-01, still being updated as of 2026-03), and the SME site clubdelapyme.com has a "free business registry lookup" tutorial (2025-10). These evergreen posts keep sending traffic over to it.
• Plugins inside enterprise software. Microsoft Power Platform has a connector for it. In March 2026 it was wired into HubSpot, so entering a company name in the CRM auto-fills registry details.
• No sign it runs ads.

Where the money comes from: block first, then figure out who's using it

For its first five years it ran on donations, and the founder's own verdict was "donations never really worked." In 2020 he quit his DevOps job at Adevinta to work on it full time. He'd long wanted to build a paid product for businesses, while the site itself stayed free.

The turn came in August 2024. He ran an experiment: searches beyond a certain number required signing up. After that, "weekly signups never stopped." Seeing who was signing up made it clear to him: "LibreBOR is no longer a niche transparency project. Thousands of businesses now use it daily to get work done, and so do public institutions. It gave them something their own government never did." In a blog post from December 2024, he wrote that he'd kept looking for new products to build without noticing that the free search box he'd started with "kept growing in usage," and that "it was the most important service."

Next came three steps:

The buyers use it for compliance checks, due diligence, CRM auto-fill, invoice verification and spotting insolvent companies, all routine daily work for businesses.

But it earns modestly. In December 2024 the founder said paying customers "just cover operating costs," and he still takes outside contract work to pay himself. He's logged more than 40 hours a week and developed neck and back problems, tendinitis and anxiety. He also keeps having to handle complaints filed with Spain's data protection agency (AEPD) asking "why is my name on your site." He hasn't published revenue since, and his LinkedIn lists the company as a 3-person operation.

The recent ranking surge: probably not real people

On the global site ranking (Tranco), it climbed from #690k to #269k over 90 days (11-05), with most of the gain concentrated in the three weeks from 2026 Aug 13 to Sep 3. But over the same period, its Spain tier and global tier in Chrome real-user data didn't move at all. Real visitors didn't increase while the ranking rose, so the likelier explanations are API call volume, AI crawlers or agents bulk-fetching its millions of pages, and more backlinks. Unverified. Its new entries into the "top 1M" in France and Italy are normal fluctuation at the tail of the list.

So this isn't a product that just "broke out." It's an established site.

Who it's competing with

Older free sites actually get more real traffic. Empresia.es and Datoscif.es take the same route (long-tail pages built from the bulletin), and both sit in Spain's "top 5 thousand" tier, one band above LibreBOR. Above them are credit bureaus like eInforma and Axesor, which sell paid reports with financials and risk scores.

More important, a batch of new APIs appeared in 2026 at a fraction of its price:

Also OpenMercantil, GetRegData, Fuentio and Boletín Claro, which target keywords with BORME tutorials. LibreBOR's parsing code was archived in April 2025, and the bulletin parsing library bormeparser is open source too. A decade ago this was hard because nobody had done it. Now you can build one in a few weeks with AI.

If it were you

C: a complete playbook to learn from, but following it in Spain doesn't pay. The Spanish niche has been crowded for 2026 years, newcomers are fighting on price, and LibreBOR itself has only reached "cost coverage" after ten years.

Who it fits: a small team of 1–2 people who understand a country's language and regulations and can handle personal data complaints, and who want to build a data site that grows traffic slowly and sells to businesses. Or a team that already has registry data cleaning capability and wants to add an output aimed at AI agents.

Who it doesn't fit: anyone who wants revenue within months; anyone planning another BORME lookup site or API in Spain (the competition already has a dozen players, and API prices have been driven to about a cent per call); anyone who is nervous about privacy complaints and legal letters.

How to do it: copy LibreBOR's skeleton, but in a different country. Find a public registry or bulletin where the government only publishes PDFs and you can't search by person. Make each entity a searchable page and collect long-tail traffic for free. Once traffic builds, copy its three-step monetization: signup wall after a search limit (which also shows you who your users are), then web freemium, then API monthly fee plus usage, and plug into enterprise software like HubSpot and Power Platform. Newcomers should ship two things from day one: an MCP interface for AI agents, and "X alternative" pages to capture the brand search of the older site. Before starting, check whether that country already has an established free site like Empresia; if it does, pick another country.

How to get customers: almost entirely search. You need many entity pages (company pages, person pages, lists of new incorporations by region), plus coverage from investigative journalists, OSINT tutorials and SME blogs, and a listing in the government open data portal's app catalog. LibreBOR has validated all of these.

Deadline: there's no holiday cutoff, but the window is closing. These AI-built low-price APIs are expanding into more countries (Prometiam already covers 36). A single-country long-tail site only has a chance if it builds a name before those players cover that country.

Signs we got it wrong: (1) LibreBOR publishes revenue clearly above "cost coverage," or raises funding or gets acquired, which would mean the model earns more than it looks and the grade should go up; (2) its Chrome real-user tier in Spain rises from "top 50k" to "top 10k," which would mean the ranking surge includes real people; (3) conversely, if a multi-country API like Prometiam covers Europe and Latin America's major countries within six months, the single-country long-tail opportunity is essentially gone.