SIAgentNews.com
← All news
DevelopingAgentsPolicy

Pandex launches API for European legal documents and AI agents

The new platform provides a structured way for AI tools to access 2.8 million official legal documents from 33 European countries.

· First reported by pandex.io

Isometric diorama: picture for the story 'Pandex launches API for European legal documents and AI agents'
AI-made isometric diorama for this story. No real people or products are shown.
2.8Mofficial legal documents held

The short version

  • Covers 2.8 million official documents from 33 of the 41 EU countries.
  • Includes 2.4 million acts and codes, 421,000 searchable passages, and 37 million citations.
  • Provides real-time access to past versions of laws and dates of effect.
  • Compatible with popular AI tools like Claude, Cursor, GitHub Copilot, and Gemini.

Why it matters

For developers and legal teams, this replaces the need to build custom scrapers for every different court or government site. It allows AI agents to provide answers grounded in actual law rather than potentially hallucinated data.

A unified source for European law

Pandex has launched a new platform designed to provide AI agents and software with direct access to official European legislation. The service aggregates 2.8 million documents from 33 of the 41 EU countries, pulling data straight from the institutions that publish them. This includes national legislation, acts, codes, and decisions from the highest courts.

The platform is built for teams that need to ensure their software handles legal information accurately. Instead of developers having to build individual scrapers for every different court or government website, Pandex provides one API and one schema across multiple jurisdictions. This structure allows for consistent data retrieval regardless of which country's laws are being queried.

The data includes 2.4 million acts and codes along with 421,000 searchable passages. As pandex.io reported, the corpus also features 37 million linked citations, which helps ground AI-generated answers in specific, verifiable legal texts. The goal is to move away from unreliable data sources toward official institutional records.

Technical integration for AI tools

The platform is designed to work with the AI tools that developers and researchers already use in their daily workflows. It offers a Model Context Protocol (MCP) server and a REST API, making it easy to connect to popular platforms like Claude, Cursor, GitHub Copilot, Windsurf, and Zed.

By offering a single API for every jurisdiction they hold, the service simplifies the engineering requirements for building legal AI products. This allows for a more streamlined way to connect software to a massive repository of legal data without needing to manage multiple different connections for different countries.

Tracking legal changes over time

One of the core features of the service is the ability to follow every change to the law across different jurisdictions. Each article in the database includes its past versions and the specific date on which each version took effect. This provides a chronological history that is essential for compliance and monitoring tasks.

The system also handles complex data hurdles that often plague legal research, such as documents hidden behind Cloudflare challenges or specific registration requirements. By providing a clean feed of data, Pandex aims to give developers a reliable way to index and build upon legal information. The platform publishes exactly what it holds and clearly notes what it does not have in its current corpus.

Detailed regional coverage

The platform includes extensive documentation from various European nations, including France, Germany, and the Netherlands. For example, it holds documents from the Bundesamt für Justiz in Germany, though some chamber decisions from 1998 to 2009 are excluded because the court refuses non-browser clients. In Austria, the platform holds documents from sources like the Bundeskanzleramt and the Verfassungsgerichtshof.

Other regions include documents from Switzerland's Bundeskanzlei and Luxembourg's Service central de législation. The platform also manages complex access issues, such as older judgments in Luxembourg that are typically hidden behind proof-of-work challenges. It covers major French institutions like the Conseil d'État, Tribunal des conflits, and Conseil constitutionnel, while noting that some French court decisions require specific accounts due to recent publication changes.

Pricing and founding opportunities

Pandex is currently in a beta phase, offering a free tier for small agents and developers. This free tier allows for 1,000 requests per day using the REST API and MCP server. For products with their own users, the pricing is based on usage, starting at €1 per 1,000 requests with no minimum monthly commitment.

For larger scale needs, the platform offers a feed starting from €1,000 a month for users who want the corpus in their own systems. There is also a special offer for the first ten companies to build on the platform, providing a fixed price of €300 a month for two years. These founding customers receive the API without counting requests, up to 1 million a month, the ability to name the next country to be loaded, and direct written answers from the people who built the system.

What to watch

  • How many new countries will be added to the 33-country coverage by the 2026 deadline?
  • How many legal tech startups will adopt the MCP server over traditional custom scrapers?

Try it

Check the Pandex institution register to see which specific court decisions or national acts are currently available for your region.

Sources and links

Read the original at pandex.io

First reported by pandex.io. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news