Lex-Ray
Making legislation legible.
Structured regulation — readable by humans, queryable by AI agents.
See it working.
01 inline_definitions
Hover any defined term — the definition comes to you, in place.
02 remote_definitions
Terms defined in other legislation resolve across instruments.
03 rapid_loading
Whole regulations, served instantly from cache.
04 keyboard_navigation
Move between provisions without touching the mouse.
05 grounded_ai
AI agents quote the law verbatim over MCP — hallucination guards built in.
06 semantic_search
Search by meaning, not just exact words.
07 version_differences
See exactly what an amendment changed — and what changed it.
08 marginalia
Pin notes, documents and commentary to any provision.
09 resources
Curated guidance and source documents — previewed instantly on hover.
Why not just use an LLM?
LLMs are great. Lex-Ray makes them better.
Lex-Ray is the structured, versioned, citable ground truth of regulation — the source your model, your agents, and you all work from. Over MCP your model reaches for Lex-Ray the way it reaches for a calculator: hand it the exact law in force, with the definitions and history in scope, and it answers from the source instead of from memory.
LLMs compress data.
A large model is lossy compression of its training data: it squeezes the corpus into weights and reconstructs a plausible version on demand, joining the dots where the detail didn't fit. Ask it for a citation and you get something shaped like a pointer to the law — but a generated citation is a guess, not a lookup: probabilistic citation, confident and fluent, with nothing connecting it to what it cites. A summary that's 95% right isn't a weaker version of the law; it's a different text.
Served exactly, or not at all.
Lex-Ray is the opposite. It keeps the source text as a rigid spine — every word, exactly as enacted — and layers structure, definitions, versions and commentary on top: never lossy, only additive. Verbatim spans at stable addresses, versions and amendments pinned, every addition labelled. The same query resolves to the same words every time; when Lex-Ray doesn't have the answer it returns nothing — it can't return something false.
Context is king.
A model is only as good as what you put in front of it. Lex-Ray is what you put in front of it — the exact legislation, the definitions in scope, the version in force, the amendments applied, fed to your model over MCP. The context that turns a plausible answer into a grounded one. Law in, law out.
Built for regulation, nothing else.
This is an interface with one job: working closely with the law. Information is layered densely on the page — definitions, versions, cross-references, your own marginalia — and it's all keyboard-navigable, searchable and quotable. You explore and surface what matters, instead of reading back whatever a single prompt returned. A different way of working — and it drives an LLM just as happily, over MCP and the API.
A professional tool.
If you're accountable for your work, anything between you and the source material has to be robust — every link from your advice back to the enacted text has to hold. Quotation and version-pinning preserve that chain; paraphrase and summary break it. Summarisation has its place: exploration, a first draft, orientation. That place is after the decision, not before it.
£0 for your first 30 days. £50/month after.
Sign up with an email. No card to start.
Firm-wide or API licensing — talk to Simmons & Simmons.
Firm-wide licensing
Rolling Lex-Ray out across your team or firm? Tell us what you need and our team will be in touch.