The EU, structured.

What you get

Quickstart

Get your first API response in under two minutes.

curl -H "Authorization: Bearer brubru_live_..." \
  "https://brubru.beresol.eu/api/v2/legislative/eur-lex/laws?q=victims+rights+directive&limit=5"

Authentication

Every /api/v1/* endpoint requires an API key. Pass it as Authorization: Bearer <key> (recommended) or X-API-Key: <key>. Keys are minted by Brubru's admin team for Professional subscribers. Rate limit is 60 requests per minute per key.

curl -H "Authorization: Bearer brubru_live_..." \
  "https://brubru.beresol.eu/api/v2/legislative/eur-lex/laws?policy_area=Trade&limit=10"

X-API-Key is also accepted for curl-friendly scripts. Bearer is the default for Postman, OpenAPI tooling, and SDKs.

Response envelope

Every paginated endpoint returns the same shape. Partners can code once against the envelope and swap datasets by changing the path.

{
  "total": 139,
  "returned": 20,
  "pages": 7,
  "page": 1,
  "limit": 20,
  "has_more": true,
  "next_page": 2,
  "remaining_pages": 6,
  "coverage_complete": true,
  "published_from": "2026-01-01",
  "published_to": "2026-01-31",
  "published_end": "2026-01-31",
  "detail_level": "Full",
  "data": [ ... ],
  "meta": {
    "source": "brubru.beresol.eu",
    "powered_by": "Brubru",
    "fetched_at": "2026-04-15T12:34:56Z"
  }
}

Code examples

Copy-paste examples for the most common integrations.

Python

import requests

API_KEY = "brubru_live_..."
BASE    = "https://brubru.beresol.eu/api/v2"

# Search laws by keyword
resp = requests.get(f"{BASE}/legislative/eur-lex/laws",
                    params={"q": "end-of-life vehicles circularity", "limit": 5},
                    headers={"Authorization": f"Bearer {API_KEY}"})
for law in resp.json()["data"]:
    print(f'{law["celex"]}  {law["title"][:80]}')

# Or browse the moat: ranked Brussels-based lobby orgs with their own news
resp = requests.get(f"{BASE}/proprietary/brussels-lobbies",
                    params={"has_news": "true", "order": "recent", "limit": 10},
                    headers={"Authorization": f"Bearer {API_KEY}"})
for org in resp.json()["data"]:
    print(org["name"], "—", org["org_type"], "—", org["last_item_date"])

JavaScript / Node.js

const API_KEY = "brubru_live_...";
const BASE    = "https://brubru.beresol.eu/api/v2";

const res = await fetch(`${BASE}/legislative/eur-lex/laws?q=end-of-life+vehicles+circularity&limit=5`, {
  headers: { Authorization: `Bearer ${API_KEY}` },
});
const { data } = await res.json();
data.forEach(law => console.log(law.celex, law.title));

MCP (any AI assistant)

Brubru exposes a Model Context Protocol (MCP) server. Connect it to Claude, ChatGPT, Gemini, Mistral, or any MCP-compatible AI agent to give your assistant live EU legislative intelligence.

// MCP server configuration
{
  "mcpServers": {
    "brubru": {
      "url": "https://brubru-production.up.railway.app/api/mcp",
      "headers": { "Authorization": "Bearer brubru_live_..." }
    }
  }
}

Client libraries

Two official Python libraries wrap the API and its EuroVoc classifier so you are one import away. Both are on PyPI and open source (MIT).

Open source on GitHub: github.com/Beresol-BV/brubru-EU-scraper-library

pip install brubru
pip install "brubru-eurovoc[local]"   # [local] adds the classifier model

brubru On PyPI

A thin Python SDK over the v2 API: typed objects over the 5-datapoint contract, automatic pagination and honest errors. The API's own client.

import brubru

bru = brubru.Client(api_key="brubru_live_...")

# Extract structured items from any EU URL, tagged with EuroVoc
items = bru.extract("https://cinea.ec.europa.eu/news_en", classify=True)

# Recent posts from a mapped entity (MEP, Commissioner, institution, journalist)
posts = bru.social.posts(entity_type="commissioner", platform="x")

eurovoc On PyPI

A standalone, open-source EuroVoc classifier that tags any text into the EU's official subject space (7,029 descriptors, 127 microthesauri, 21 domains), multilingual including Catalan. A modern successor to the 2021 PyEuroVoc package, and the natural home for a retrained long-context model.

import eurovoc

eurovoc.classify("Markets in crypto-assets regulation")
# -> [{"descriptor": "financial instrument", "mt": "2426",
#      "domain": "24", "score": 0.88}, ...]

# or run it through the hosted classifier over the API
eurovoc.classify("...", backend="brubru", api_key="brubru_live_...")

Full documentation

The full API reference lives on its own microsite: introduction, authentication, response envelope, errors, rate limits, every endpoint documented with parameters and curl examples, a matrix of every EU institution we ingest from, a glossary for every acronym, role-specific use cases, and a dated changelog.

Rate limits and errors

Error responses use a canonical envelope: { error, reason_code, request_id }. Every response also carries an X-Request-Id header for support traceability.

Start building