ranking-empresas.eleconomista.es logo
ranking-empresas.eleconomista.es

search-company-data

Installation

Adds this website's skill for your agents

 

Summary

Given a company name, resolve it against the eleconomista company ranking directory and return its fiscal ID, activity, address, financials, and ranking positions — read-only.

FIG. 01
FIG. 02
FIG. 03
FIG. 04
FIG. 05
SKILL.md
178 lines

Search Spanish Company Data by Name

Purpose

Given a Spanish company name, this skill resolves it to a canonical company profile on ranking-empresas.eleconomista.es (the elEconomista / INFORMA D&B company-ranking directory) and returns structured data about that company: fiscal ID (NIF/CIF), legal name, activity/CNAE, registered address, phone, website, legal form, headline financials (sales, total assets, EBITDA, result, capital), and its national / provincial / sector ranking positions. It is strictly read-only — it only reads public directory pages, never submits forms or logs in.

When to Use

  • You have a company name (exact or partial) and need its fiscal ID, sector, or registered address.
  • You need headline financials (annual sales / turnover, total assets) for a Spanish company from the INFORMA D&B dataset.
  • You want a company's position in the national, provincial, or sectoral revenue ranking.
  • You need to disambiguate between several similarly-named Spanish entities (e.g. all the SANTANDER * subsidiaries) before drilling into one.

Workflow

The site exposes everything needed over plain HTTP with no anti-bot protection (homepage probe returned 200, no CAPTCHA / WAF). The optimal path is two cheap fetches — do not script a browser.

Step 1 — Resolve the name to a canonical company (predictive-search API)

GET the internal predictive-search endpoint with the raw name as query:

https://ranking-empresas.eleconomista.es/servlet/app/portal/EMP/prod/XML_AJAX_BUSQUEDA_KEY/?query={url-encoded name}

It returns an XML <keywords> document with up to ~10 <keyword> matches, each containing:

  • <nombreMostrar> — display name (e.g. INDITEX LOGISTICA SA)
  • <nombreNorm> — normalized name used to build the detail-page URL (e.g. INDITEX LOGISTICA)
  • <esEmpresa> — 1 when the row is a company

Pick the best match. Always use <nombreNorm> (not <nombreMostrar>) to build the URL — the normalization is lossy and inconsistent (drops/keeps SA/SL/SLU unpredictably, strips accents and punctuation), so never hand-derive the slug from the display name.

If there are zero <keyword> elements, the name has no match → return success: false.

Step 2 — Fetch the detail page and parse it

Build the detail URL from <nombreNorm> by replacing spaces with -, URL-encoding, and appending .html (relative to the site root):

https://ranking-empresas.eleconomista.es/{NOMBRE-NORM-WITH-DASHES}.html
# e.g. "INDITEX LOGISTICA" -> https://ranking-empresas.eleconomista.es/INDITEX-LOGISTICA.html

GET that page and extract data from two sources:

  1. JSON-LD (<script type="application/ld+json">, @type: LocalBusiness) — cleanest source for name, taxID (the NIF/CIF), isicV4 (CNAE code + label), telephone, description (objeto social), and structured address (streetAddress, addressLocality, addressRegion, postalCode) + geo lat/long.
  2. Rendered HTML body — for fields not in JSON-LD: Forma Jurídica (legal form), Capital Social, Página Web, Actividad, latest-year Ventas (sales) and Total Activo (total assets), Ebitda, Resultado, number of employees, and the three ranking positions (Ranking Nacional Posición, Ranking Provincial Posición … de {province}, Ranking Sectorial Posición … de {sector}).

Money values render in Spanish format (140.475.130 € = 140,475,130). Strip . thousands separators and the € before casting to a number.

Step 3 — Emit the JSON

Return the object shaped like the ## Expected Output schema below, echoing the input query, the chosen company, and the other <keyword> display names as candidates so a caller can re-run against a different match if the auto-pick was wrong.

Browser fallback (only if the fetch path is unavailable)

If you must drive a real browser (e.g. the internal endpoint changes):

  1. browse open https://ranking-empresas.eleconomista.es/ --remote (bare session — no --proxies/--verified needed).
  2. browse fill "#search-nopred" "{name}" then browse press Enter. This lands on a static results page https://ranking-empresas.eleconomista.es/busqueda_rankings_{QUERY-NORM}.html listing matching companies, each linking to a …/{NAME}.html detail page.
  3. Do not rely on clicking the result links — in testing they did not navigate on click. Read the href of the desired result and browse open that URL directly.
  4. On the detail page, browse get text body (or read the JSON-LD) and parse as in Step 2.

Site-Specific Gotchas

  • No anti-bot. A bare Browserbase/fetch session works; --proxies/--verified are unnecessary. The successful run used neither.
  • The detail-page slug comes from <nombreNorm>, not the display name. Normalization is lossy and not a simple uppercase+dash of the display name: SANTANDER INSURANCE SL. → SANTANDER INSURANCE (suffix dropped) but SANTANDER GLOBAL TECHNOLOGY AND OPERATIONS SLU keeps SLU. Accents are stripped (ESPAÑA → ESPANA). Always take the slug from the search API's <nombreNorm>, replace spaces with -, then URL-encode and append .html.
  • Predictive endpoint only returns matches for queries longer than 3 characters (the client waits until length > 3 and debounces ~500 ms). Very short queries may return an empty <keywords> even for real companies — pass the fullest name fragment you have.
  • Empty result shape: a no-match query still returns HTTP 200 with a <keywords> element containing <textoBuscadoOriginal>/<textoBuscadoNormalizadoURL> but zero <keyword> children. Treat "no <keyword>" as not-found, not as an error.
  • JSON-LD is the reliable NIF/CIF source. The visible HTML obfuscates the fiscal ID (there is a nif-encode="…" attribute rather than plain text); read taxID from the JSON-LD LocalBusiness block instead of scraping it from the body.
  • Financials are latest-available-year and Spanish-formatted. The page shows values like Total Activo 2025 136.214.685 €; the year label varies by company and data vintage. Strip . and €; capture the year alongside the figure if the caller needs it.
  • Data provenance: figures come from INFORMA D&B S.A.U. The page pushes an "Informe Ampliado" (extended report) behind an eInforma registration — ignore those CTAs; all fields this skill returns are on the free public page.
  • A static results-page fast-path also exists: https://ranking-empresas.eleconomista.es/busqueda_rankings_{QUERY-NORM}.html renders the full match list server-side (useful if you prefer HTML scraping over the XML API), where {QUERY-NORM} is the textoBuscadoNormalizadoURL value the API returns.
  • browse snapshot noise: the CLI in the sandbox prints an "Update available" banner to stdout that can corrupt snapshot/accessibility-tree parsing. Prefer browse get text body / browse get markdown body (or the JSON-LD) over browse snapshot for extraction on this site.
  • Ambiguity is common. Names like Santander, Inditex, Iberdrola return many subsidiaries. Return the full candidate list and pick the top match (usually the parent / highest-ranked entity) unless the caller specifies otherwise; the assumption made here is "first esEmpresa=1 match" as the default.

Expected Output

Successful resolution:

{
  "success": true,
  "query": "Inditex",
  "company": {
    "name": "Inditex Logistica Sa",
    "tax_id": "A70268669",
    "activity": "Depósito y almacenamiento",
    "cnae": "5210",
    "legal_form": "Sociedad anónima unipersonal",
    "objeto_social": "Almacenamiento en general. Transporte y distribución de mercancías.",
    "address": "Avenida Deputacion, S/N",
    "locality": "Arteixo",
    "region": "Coruña",
    "postal_code": "15142",
    "geo": { "lat": 43.3166768, "lng": -8.5014506 },
    "phone": "981185400",
    "website": "www.inditex.com",
    "capital_social_eur": 100000,
    "sales_eur": 140475130,
    "total_assets_eur": 136214685,
    "ebitda_eur": 62497880,
    "result_eur": 33284252,
    "financials_year": 2025,
    "ranking_national": 2092,
    "ranking_provincial": { "position": 62, "province": "Coruña" },
    "ranking_sector": { "position": 11, "sector": "Depósito y almacenamiento" },
    "detail_url": "https://ranking-empresas.eleconomista.es/INDITEX-LOGISTICA.html"
  },
  "candidates": ["INDITEX LOGISTICA SA"],
  "error_reasoning": null
}

Ambiguous name (several matches — first is auto-picked, rest returned for the caller):

{
  "success": true,
  "query": "Santander",
  "company": { "name": "Santander Global Technology And Operations Slu", "detail_url": "https://ranking-empresas.eleconomista.es/SANTANDER-GLOBAL-TECHNOLOGY-AND-OPERATIONS-SLU.html", "tax_id": "..." },
  "candidates": [
    "SANTANDER GLOBAL TECHNOLOGY AND OPERATIONS SLU",
    "SANTANDER INSURANCE SL.",
    "SANTANDER BACK-OFFICES GLOBALES MAYORISTAS SA",
    "SANTANDER MEDIACION OPERADOR DE BANCA SEGUROS VINCULADO SA",
    "SANTANDER FACILITY MANAGEMENT ESPAÑA SL."
  ],
  "error_reasoning": null
}

No match:

{
  "success": false,
  "query": "zzznotarealcompany",
  "company": null,
  "candidates": [],
  "error_reasoning": "Predictive-search endpoint returned a <keywords> document with no <keyword> matches for this name."
}
Search Spanish Company Data by Name · browse.sh