About

What's in the knowledge graph

The TCLP knowledge graph holds The Chancery Lane Project's curated, climate-aligned legal content:

• Clauses — model contract clauses drafted to reduce greenhouse gas emissions and embed climate commitments in agreements.
• Glossary terms — defined concepts (for example, scope 3 emissions) that recur across the clauses and guides.
• Guides — practical explainers that put the clauses and terms in context.

Everything served by the connector comes from this knowledge graph. The content is curated by TCLP and its contributor community, not generated on demand.

The tools

The connector exposes four tools. Each answers a different kind of question:

Tool	Use it when you want to…
search	Find relevance-ranked results across all content for a free-text query.
entity_lookup	Start from a named concept (e.g. "scope 3 emissions") and find the clauses and terms that mention it.
taxonomy_facets	Discover the available categories — facet names and their value slugs (sector, practice area, application, jurisdiction, and so on).
taxonomy_content	Browse content filtered by one or more taxonomy values, rather than by relevance.

A typical browse uses the last two together: call taxonomy_facets first to see which slugs exist, then taxonomy_content with the slugs you want.

How search works

A search here doesn't just match a query against text. The connector searches three layers at once:

• The exact words in your query.
• The semantic meaning behind those words, so closely related content surfaces even when the wording differs.
• The relationships between concepts captured in the knowledge graph.

The results are fused into a single ranking. So a search for a single term doesn't just return a list of documents — it surfaces the relevant clauses, the practical guidance, the supporting definitions, and the wider web of legal and climate concepts that connect them. The answer tells you not only what a concept is, but how it relates to the broader landscape of climate-aligned legal practice.

Provenance

All content is drawn from The Chancery Lane Project's knowledge graph. The connector is a read-only window onto that data — it does not author, summarise, or alter content.

Design principles

The connector and the API behind it are built to a few deliberate constraints:

• Read-only. It searches and returns content. It cannot create, update, or delete anything.
• No LLM dependency in the API. Search and ranking are deterministic; the API does not call a language model. Any reasoning over results happens in your assistant, not on the server.
• No raw query language exposed. There is no open Cypher or GraphQL endpoint; complex querying stays on the client side, behind the well-defined tools above.