MCP App Store
by KeenEthics
Insuraviews icon

Insuraviews MCP Server for Claude and ChatGPT

By Insuraviews

Overview

Insuraviews gives insurance product, pricing and actuarial teams fast access to P&C rate and rule filings from carriers across the US. With the Insuraviews connector, Claude can search filings, pull the details of a specific filing, read its supporting documents, and identify carriers and groups, so you can answer competitive questions in conversation instead of digging through filing portals. Key capabilities: - Search filings by carrier, state, line of business, filing type and date - Retrieve filing details, including rate changes, effective dates and status - Open filing attachments such as rate exhibits, actuarial memoranda and rule manuals - Resolve carrier and group names to the right entities - Count and trend filing activity across markets Example uses: summarize a competitor’s latest homeowners rate filing in Texas, compare personal auto rate changes across top writers in a state, or track filing activity in a line over the past year. Access requires a license to the Insuraviews Data Connector. Contact Insuraviews at sales@insuraviews.com to add it to your subscription.

How to connect

  1. Claude: Open the Insuraviews listing in the Claude connectors directory and choose Connect. Claude listing Insuraviews exposes 6 tools here.
  2. ChatGPT: Open the Insuraviews listing in the ChatGPT app directory and choose Connect. ChatGPT listing Insuraviews exposes 5 tools here.

Tools

count_filings

ChatGPT
Count US property and casualty insurance regulatory filings (SERFF) matching filters and an optional keyword, in total or grouped by up to two of: state, year, month, carrier, company, lineOfBusiness, dispositionStatus. Use it for how many, trends over time, which carriers or states file most, and share by state: one call instead of paging search_filings or calling it with pageSize 1 to tally. Then call search_filings for the filings behind a group. Takes the same filters and query as search_filings (carriers and companies by name or id); dateType picks the date that year, month, dateFrom and dateTo use. With query, counts are of filings with at least one matching passage, not of every filing on the topic, and are estimates (totalFilingsIsApproximate). Without query, counts are exact. Groups come largest first, or in time order when the first groupBy is year or month; up to 500 per call (maxGroups), and groupsTotal says how many there are. Example: {"query": "\"water score\"", "linesOfBusiness": ["Homeowners"], "dateFrom": "2021-01-01", "groupBy": ["year", "state"]}

get_attachment

ChatGPT
Reads one attachment of a SERFF filing (a markup, form, rate exhibit, memo or objection letter) as its own text, page by page. Use it to check specific facts before stating them: which companies a change applies to, page numbers, exact definition wording, rates and factors. AI-written summaries can get these wrong. Get key from get_filing with sections ["attachments"], or from a search_filings match. Limits: up to 25 pages per call (maxPages, default 10) and about 40,000 characters, whole pages only (a single longer page is cut, marked [truncated]); continue from nextPage when present. Pages with no extracted text are skipped; original files are not available. The text is extracted from PDFs and can be garbled, especially in redlined markups; say so when it is, rather than guessing. Example: {"serffTrackingNumber": "ABCD-134000000", "key": "<key from get_filing>"}

get_filing

ChatGPT
Full detail for one SERFF filing: who filed it, what it changes, the rate impact by company, and how the regulator disposed of it. Use after search_filings, or when the user names a SERFF number. By default returns the summary (AI-written), rate impact and disposition; pass company to get only that company's rate impact; add rating_factors, objections or attachments to sections only when the question needs them. Attachments list each document's key (for get_attachment, which reads its text) and an AI-written executiveSummary. Limits: each long text field (summary, detailed summary, change log, objection summary) is cut at 12,000 characters, marked [truncated: N more characters]; rating factors list the first 150 (ratingFactorCount gives the total); attachments are all listed, but their text comes only from get_attachment. Original PDF files are not available. Cite each filing by its SERFF number linked to citation.url (the filing in the Insuraviews app), e.g. PRGS-135049001. Example: {"serffTrackingNumber": "ABCD-134000000", "sections": ["summary", "objections"]}

resolve_entity

ChatGPT
Turns a name the user typed into the exact value search_filings filters on, closest matches first. Carriers and companies return an id in value: pass it as carrierIds or companyIds. Other kinds return the exact spelling to pass as that filter. Call it before filtering by carrier or company, and whenever a filter value might not match the data's spelling. A similarity near 1 is a confident match; below about 0.3, ask the user which one they mean. States are stored as full names (Texas); abbreviations (TX) are accepted. Limits: up to 20 matches (limit, default 5). Example: {"kind": "carrier", "text": "state farm"}

search_filings

ChatGPT
Search US property and casualty insurance regulatory filings (SERFF) by keyword and filters. With query, matches filing text (filings, summaries, objections, rating factors, attachments) and returns matching excerpts with «matched words» marked, most relevant first. Every bare word must appear in the same passage, so use a few distinctive words or a quoted phrase, not a sentence; put OR (uppercase) between alternatives. Without query, lists filings matching the filters, newest first by dateType (submission date by default); rows are one per filing and company, and totalFilings counts filings. For approved or disposed filings, set dateType to dispositionDate. Lines are broad (Personal Auto includes motorcycle and RV; businessowners sits under CMP); narrow with subTypesOfInsurance from resolve_entity, e.g. 19.0001 Private Passenger Auto (PPA). Carriers and companies can be given by name (carriers, companies); resolved in the result says what each name matched. An unclear name returns its closest candidates instead: ask the user or pass ids. Use resolve_entity to browse names, for brands that aren't carriers (GEICO), and to check other filter values, which must match the data's spelling. Limits: up to 25 filings per call (default 10); use page for more, or narrow the filters when there are thousands. Keyword results show up to 3 matching passages per filing, 2 excerpts each of up to 800 characters, shortened (excerptsTrimmed) when a page would pass about 50,000 characters; read the full text with get_attachment. Both modes include each filing's rate impact and dates (keyword results list up to 5 companies per filing), so call get_filing only for summaries, objections, rating factors or attachments. totalFilings counts up to 10,000; totalFilingsIsLowerBound says when there are more. Cite each filing by its SERFF number linked to citation.url (the filing in the Insuraviews app), e.g. PRGS-135049001. Example: {"query": "\"roof surfacing\"", "states": ["Texas"], "linesOfBusiness": ["Homeowners"], "dateFrom": "2026-01-01"}

count_filings

Claude

get_attachment

Claude

get_filing

Claude

ping

Claude

resolve_entity

Claude

search_filings

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/insuraviews"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/insuraviews)