The Chunk Model
Each regulation is split into chunks — topic-sized pieces that answer specific questions. A chunk is typically 500-1500 words and covers one concept.
Example: UK GDPR is split into chunks like lawful-basis,
consent, breach-notification, etc.
Content Layers
Each chunk can have multiple layers:
Summary Layer
Plain English explanation. Answers "what do I need to know?" Written for quick comprehension by agents and humans.
Source Text Layer
Verbatim legal text from the regulation. For regulations like the EU AI Act, we provide full article text alongside summaries.
Citations
Every claim in our content includes a citation to the source law. This lets agents verify information and provide references to users.
| Regulation Type | Citation Format | Example |
|---|---|---|
| UK GDPR | [Art X] | [Art 6(1)(a)] |
| UK Statutes | [s.XX] | [s.21(2)] |
| FCA Handbook | [COBS X.X] | [COBS 4.12A.1R] |
| EU Regulations | [Art X] | [Art 5], [Annex III] |
URL Structure
/regulations/[jurisdiction]/[regulation]/[chunk]
Examples:
/regulations/uk/gdpr/lawful-basis
/regulations/uk/fca-financial-promotions/risk-warnings
/regulations/eu/ai-act/articles/art-5Accessing Content
Every page is available in multiple formats:
/regulations/uk/gdpr/consent— HTML for browsers/regulations/uk/gdpr/consent.md— Markdown for agents
The markdown version includes frontmatter with metadata (title, citations, parent regulation, source URLs).
Regulation Index
Each regulation has an index.md that maps all its chunks.
The llms.txt file provides topic-based routing for agents.