MCP App Store
by KeenEthics

Overview

New Zealand government policy change. See it early. Understand it clearly. Act while it matters. Focused exclusively on Aotearoa New Zealand, Lumini is a government and public policy platform built on two foundations: 1. A comprehensive, continuously updated collection of New Zealand government, parliamentary, legislative and regulatory information. 2. Practical tools that help organisations discover and monitor developments, investigate issues, understand their implications, prepare effective responses and pursue policy, legislative and regulatory settings that better reflect their needs. The read-only Lumini MCP connection provides authorised users access to our source-linked New Zealand government and policy dataset, specialist monitoring, and policy and advocacy workflows directly within Claude.

Tools

get_consultation_deadlines

ChatGPT
List upcoming consultation deadlines sorted by closing date. Tip: use for submission planning — "what closes soon?" e.g. "what government consultations close in the next two weeks?"

get_legislation_summary

ChatGPT
Summarize legislative activity — bills by stage, recent activity, upcoming deadlines — then layer the caller's organisation over it: each recent bill's enrichment (bill_summary + impact_analysis), whether it is already tracked, the org's context, and guidance to recommend which untracked bills the org should track and why. Tip: the go-to "what's moving in Parliament, and what should we be tracking?" overview, and a strong first step in a weekly briefing. e.g. "what bills progressed in the last 7 days, and which should we be tracking?"

get_minister_report

ChatGPT
Generate a minister report — their portfolio, positions, sponsored bills, voting record, key statements, and what they advocate for and against. Tip: use to profile an MP or minister. Returns source material (bills, Hansard, excerpts) for you to synthesise. e.g. "what does Minister Simeon Brown advocate for and against?"

get_my_context

ChatGPT
Returns the signed-in user (name, email), their access tier, and their organisation (name, sector, summary, website). Identity and organisation are resolved from the authenticated session, so this information is available without asking the user. Tip: a good first call in a conversation, so the user and their organisation can be addressed by name. e.g. "who am I and what org am I in?"

get_usage_stats

ChatGPT
Platform usage analytics for an organisation — chat counts, document counts, active users, a breakdown of alert subscriptions (keyword / portfolio / targeted / rapid / select-committee), reports generated, and bills being tracked. Scoped to the signed-in user's organisation automatically. Tip: use for "how are we using Lumini?" questions. e.g. "how many active users, reports and tracked bills did we have this month, and what kind of alerts are set up?"

get_weekly_top_10

ChatGPT
Personalised top 10 most relevant items this week — based on the user's alert subscriptions, tracked bills, and organisation sector, scoped to the signed-in user's organisation automatically. Tip: use for "what should I focus on this week?" Returns relevance-ranked context — documents, bill activity, and consultations — that can be summarised into a top-10. e.g. "give me my weekly highlights".

search_committee_hearings

ChatGPT
Deep search over select committee hearing transcripts — full-text over what was said, plus filters for committee, bill, and who attended (witness name/organisation). Transcripts are available from 1 May 2026 onwards; hearings before that date will not appear even if a bill or committee is otherwise active. Note: witness filters match who ATTENDED a hearing (structured attendee records), not who said a particular thing — the transcript itself has no per-speaker attribution, so search matches anywhere in the transcript regardless of who was speaking. Tip: pass bill_title as a name or number ("water services bill", "127-2") rather than a uuid — it is resolved fuzzily server-side, and the response reports what it matched so a wrong resolution is visible. committee is matched as a substring ("health" matches "Health Committee"), so a close-enough guess is fine. e.g. "find hearings on the Water Services Bill where MBIE gave evidence".

search_documents

ChatGPT
Hybrid search across the Lumini document corpus — fuses semantic (vector) similarity with keyword (full-text) matching and an optional recency bias. Scoped to the signed-in user's organisation automatically. Tip: use for "find / what does X say about / dig up" questions over ingested documents (Hansard, legislation text, policy, consultations). The keyword signal makes it reliable for exact terms vector search alone misses — proper nouns, acronyms, bill names, section numbers, te reo Māori. Lower threshold for broader semantic results; raise it for only close matches. Use after_date/before_date for time-bounded questions ("anything since March", "in 2025"), and doc_id with scope: "document" to search within a single document. e.g. "search our documents for the Three Waters reform".

get_consultation_deadlines

Claude

get_legislation_summary

Claude

get_minister_report

Claude

get_my_context

Claude

get_usage_stats

Claude

get_weekly_top_10

Claude

search_committee_hearings

Claude

search_documents

Claude

Similar Apps