MCP Connector · Claude · ChatGPT · Perplexity · Gemini

Query verified luxury-asset prices from your AI.

The ALT/FNDATA MCP connector puts 13M+ verified auction and private-sale results inside any MCP-capable assistant. Ask in plain language and get answers grounded in real cleared transactions, each with a link back to the source lot.

Works in ClaudeChatGPTPerplexityGemini
Get your API key Read the docs

Server URL: https://mcp.altfndata.com/mcp

See it in your AI
YouUsing ALT/FNDATA, show the ten highest prices for Patek Philippe watches sold in 2024, with the auction house and date.
Your assistant calls query_table(table: "watches", filters: sold & designer = Patek Philippe & 2024, sort: usd_price_decimal desc, limit: 10)
**10 results** from `watches` 1. Patek Philippe, Ref. 1518 perpetual calendar chronograph vendor: Christie's sale_date: 2024-11-05 usd_price_decimal: 3,860,000 2. Patek Philippe, Ref. 2499 second series vendor: Sotheby's sale_date: 2024-05-13 usd_price_decimal: 3,200,000 ... eight more, sorted by price

Illustrative shape. Your assistant reads the schema, runs one read-only query, and answers from real sold-lot rows.

Overview

One connector. Real cleared prices. Any AI.

The Model Context Protocol (MCP) is an open standard for connecting AI assistants to external data. The ALT/FNDATA MCP connector exposes our cleared auction and private-sale dataset, 13M+ realized results, to any MCP-capable assistant. You sign in once with your ALT/FNDATA account over OAuth, then ask questions in natural language.

Every row the connector returns is a real transaction with a link back to the source lot, not an asking price and not an estimate. The assistant reads the schema, runs a read-only query, and answers from the actual sold rows. Because the connector is read-only, it can look up prices, run statistics, and check values, but it never changes your data or places any order.

The short versionGet an ALT/FNDATA key, add https://mcp.altfndata.com/mcp as a connector in your assistant, sign in, and ask about verified prices across fifteen luxury and collectible categories.

The connector

Nine read-only tools.

Grouped by what they do. Your assistant picks the right one for each question; you just ask in plain language.

Explore

Find the data and pull rows

list_tablesLists the fifteen category tables you can query.
get_table_schemaReturns the columns and types for a table so the assistant knows what it can filter and sort on.
query_tableRuns a read-only query against one table with filters, sorting, and limits.
searchWhole-market keyword search that fans out across every category at once.
Analyze

Turn rows into a read on the market

market_statsCount, median, percentiles, and top vendors for a filtered slice of the market.
price_trendRealized price over time, so you can see where a segment is heading.
market_indexThe headline ALT/FNDATA index for watches, handbags, and cars.
Value

Put a number on a single item

price_checkA single-item value range with a verdict and comparable sales, each with a source link.
Monitor

See what you are watching

list_alertsReads back the market alerts on your account. The connector authenticates with OAuth or an API key, so alerts are set in the ALT/FNDATA app, not created through the connector.

The data

Fifteen categories, one schema.

One connector, one consistent schema, fifteen categories of verified cleared results. Note the deliberate split between jewelry (finished, mounted pieces) and gems (loose, unmounted stones), because a mounted ruby ring and a loose ruby trade in different markets.

WatchesWristwatches and pocket watches
JewelryFinished, mounted pieces
GemsLoose, unmounted stones
HandbagsDesigner and collectible bags
CarsClassic and collector automobiles
MotorcyclesCollector and vintage bikes
AircraftPrivate jets and aircraft
Wine & whiskyFine wine and rare spirits
Works of artDecorative and applied arts
Fine artPaintings, prints, sculpture
DesignCollectible design objects
FashionCouture and collectible apparel
ShoesCollectible and rare footwear
AutomobiliaMotoring memorabilia and parts
BooksRare books and manuscripts

Prices are normalized to USD so slices compare cleanly across houses and currencies, and every result keeps the vendor, the sale date, and a link back to the original lot.

Integrations

Works in four assistants.

The connector is a standard remote MCP server, so it works anywhere custom remote connectors are supported. Availability and setup differ by platform; the four cards below cover each one.

Claude

Easiest · Free and up

Works in Claude web, Claude Desktop, and Claude Code. Custom connectors are available on Free and up; the Free plan is limited to one custom connector.

  1. Open Settings, then Connectors.
  2. Choose Add custom connector.
  3. Paste https://mcp.altfndata.com/mcp.
  4. Click Connect and sign in with your ALT/FNDATA account.

The full set of read tools becomes available as soon as you are signed in.

ChatGPT

Plus and up · Dev Mode

Added through Developer Mode on the web, available on the Plus, Pro, Business, Enterprise, and Edu plans. Gives the full set of read tools.

  1. Open Settings, then Connectors or Apps, then Advanced, and turn on Developer mode.
  2. Add a new MCP server and enter https://mcp.altfndata.com/mcp.
  3. Authorize, then sign in with your ALT/FNDATA account.

Developer Mode surfaces custom MCP servers in the ChatGPT web app.

Perplexity

Perplexity Enterprise

Custom remote connectors are available on Perplexity Enterprise today, with broader rollout coming.

  1. Open Settings, then Connectors.
  2. Add the URL https://mcp.altfndata.com/mcp.
  3. Authorize, then sign in with your ALT/FNDATA account.

Once authorized, Perplexity can call the connector inside your research threads.

Gemini

Enterprise & API

Gemini Enterprise supports custom remote MCP servers over OAuth as a data store, and the Gemini API (Managed Agents) supports remote MCP for developers. The consumer Gemini app does not expose a custom-connector interface yet.

  1. In Gemini Enterprise, add https://mcp.altfndata.com/mcp as a custom remote MCP data store and authorize over OAuth.
  2. Or, with the Gemini API, register the remote MCP server in your Managed Agent configuration.
  3. Sign in with your ALT/FNDATA account when prompted.

Gemini support is Enterprise and API for now; the consumer app cannot add custom connectors.

The same across every platform. The endpoint is always https://mcp.altfndata.com/mcp, you sign in with your ALT/FNDATA account over OAuth, and a paid ALT/FNDATA plan is required to return data. The free tier returns a 402 until you upgrade.

10 Ways to Use ALT/FNDATA in Your AI

For appraisers

A basis of value that survives an audit.

Pull documented, cleared comparables for any item, straight from your AI, with the vendor and date behind every line.

  • Real sold lots, not asking prices
  • A citable line for every comp
  • Comp sets that match the piece
  • A read on predictability
See the appraisers page →

Use cases

What people use ALT/FNDATA in their AI for

Family office

Size and monitor a collection at the market level, and see how liquid each category is.

Hedge-fund analyst

Read auction demand as an alternative signal on a listed brand, tied to its parent ticker.

Wealth advisor

Put a defensible, sourced market range next to a client's passion asset.

Brand strategist

Benchmark a house's resale strength against rivals and spot where demand is rotating.

Appraiser

Pull a clean, dated, sourced comp set into a valuation file in minutes.

Financial journalist

Stand up or knock down a market-trend claim with sold-lot evidence.

Dealer and trader

Price a specific reference and time when to hold or move a segment.

Insurance and risk

Set scheduled values and test claimed values against recent comparable sales.

Go deeper

Get started

Get Your API Key

Create an ALT/FNDATA account, choose a plan, and get the key that signs you in to the connector.

Open the app → Technical Docs

Full Developer Docs

The connector exactly as it ships: tool parameters, return shapes, auth, plan limits, and error codes.

Read the docs → Use-Case Playbook

Prompts by Persona

Eight roles and more than twenty copy-paste recipes, each with its output shape and the caveat it carries.

Explore the playbook → Coverage & Methodology

How to Trust an Answer

What the data covers, how fresh it is, USD normalization, and the recency caveat, in full.

Read the methodology →