analyze_document_relevance, find_person_in_document and find_party_in_document all exist to avoid loading a full document: the first reports what is in it before you commit, the other two return positions by fuzzy match, and get_document_content then takes an offset so you can read around a hit rather than from the top. The scope is one parliament — every tool addresses Tweede Kamer material, and the directory, committee and voting tools carry no historical membership, only what is current.
A research layer over the Dutch Tweede Kamer's documents, activities, votes and member directory, with tools for reading inside a document as well as finding it.
- Full-text search across documents, activities and cases, filterable to one of those types or to a category such as moties or vragen, with quoted phrases and NOT exclusion supported
- Voting results per motion or bill: date, submitter, accepted or rejected, for and against counts, and which parties voted which way
- A directory of sitting MPs with IDs, parties and faction memberships, their official portraits, and today's birthdays
- Committees with their members, leadership roles and recent work, plus the upcoming debates and meetings sorted by date
- Document metadata and links to the PDF and the official page, and the document text itself either paged or centred on a byte offset
- Locations of a person or a party inside a document by fuzzy match, and a relevance pre-check that reports what a document contains before you read it
Nothing to supply — no account and no key. Documents are addressed by the numbers the search tools return, so a lookup starts from a search rather than from an identifier you already hold.
One command — npx -y @r-huijts/opentk-mcp