MCP reference
fetch · ChatGPT
Read a search result at the ChatGPT endpoint.
Call
Required argument id: copy the document ID exactly from search. The ID includes document-type information; do not remove or alter any part.
{
"id": "<id from search>"
}
Response
fetch returns id, title, text, url and, where available, metadata. An available AI summary is labelled and placed before the original text, including facts and legal issues with any available supporting references.
Official paragraph numbers appear as [Rn. 42]. If a document exceeds the output limit, the response includes a truncation notice and a link to the full text on LawDB. This endpoint does not provide a paging cursor.
If the document is still being loaded from its source, the response states a waiting time. Repeat retrieval afterwards using the same id.