Finmap.org MCP Server Historical Stock Data

Overview

The Finmap.org MCP server exposes historical market data from multiple exchanges (US, UK, Russia, Turkey and Hong Kong) using the Model Context Protocol (MCP). It is designed to feed AI agents, tools and integrations with structured time series, sector breakdowns, company profiles and aggregated metrics such as market capitalization, volume and trade counts. The service also provides ready-made visualization payloads (treemaps and histograms) usable by downstream clients.

For developers building research assistants, trading analytics, or GPT-powered workflows, the MCP server acts as a lightweight, queryable data layer. You can use the hosted endpoint for rapid prototyping or run the npm package locally for lower latency and offline access. The API is small and predictable, suitable for embedding in agents, Claude/other desktop clients, or server-side automation.

Features

• Historical price & volume time series across multiple exchanges
• Sector listings and sector-level aggregates (counts, value, market cap)
• Company profiles and per-ticker snapshots (market cap, last price, trades)
• Data visualizations ready for front-ends: treemaps and histograms
• Compact HTTP API and MCP tool definitions for agent integrations
• Hosted endpoint for zero-install use + an npm package for local deployment
• Frequent updates (exchange-dependent cadence from minutes to months)

Supported Exchanges (summary)

Installation / Configuration

Option A — Use the hosted server (no install)

Server URL: https://mcp.finmap.org

Example: configure a Claude Desktop MCP remote

{
  "mcpServers": {
    "finmap": {
      "command": "npx",
      "args": ["mcp-remote", "https://mcp.finmap.org"]
    }
  }
}

Option B — Run locally via npm

# Install globally
npm install -g finmap-mcp

# Or run on-demand with npx
npx finmap-mcp

Example local MCP config for Claude Desktop

{
  "mcpServers": {
    "finmap": {
      "command": "npx",
      "args": ["-y", "finmap-mcp"]
    }
  }
}

Note: running locally can reduce latency and allow offline queries against the packaged dataset.

Available Tools / API Endpoints

The server exposes both MCP tools and a compact HTTP API suitable for GPT Actions or direct use.

Common MCP tools (names you can call from an agent):

• list_supported_exchanges — metadata for every supported exchange (id, name, country, currency, earliest date, update frequency)
• list_exchange_sectors — list business sectors on a given exchange and date, with company counts
• list_sector_companies — list tickers in a sector (with summary metrics)
• search_companies — fuzzy/company search by name or ticker
• market_overview — aggregate market metrics and visuals (treemap/histogram)
• sector_performance — time-series performance aggregated by sector
• rank_stocks — top/bottom ranking by market cap, volume, or returns
• stock_snapshot — recent snapshot for a single ticker
• company_profile — static company metadata and description

HTTP API endpoints:

• GET /api/openapi.json
• GET /api/list-exchanges
• POST /api/list-sectors
• POST /api/list-sector-companies
• POST /api/search-companies
• POST /api/market-overview
• POST /api/sector-performance
• POST /api/rank-stocks
• POST /api/stock-snapshot
• POST /api/company-profile

Quick curl example — list sectors for LSE on a date

curl -X POST "https://mcp.finmap.org/api/list-sectors" \
  -H "Content-Type: application/json" \
  -d '{"exchange":"lse","date":"2025-03-01"}'

A GPT Actions schema (gpt-actions.yaml) is provided in the repo for importing into agents that support the Actions format.

Use Cases

• Building an AI research assistant that answers questions like “Which LSE sectors gained the most market cap in the last month?” — combine sector_performance and market_overview.
• Powering a dashboard: fetch market_overview to render treemaps by sector and histograms of trade volume distribution.
• Event-driven alerts: an agent periodically calls rank_stocks to detect sudden market-cap changes or unusual volume for a watchlist.
• Company research: use search_companies + company_profile + stock_snapshot to assemble a one-page summary for a ticker (profile, recent trading stats, market cap).
• Backtesting feature engineering: pull historical series by exchange and apply transformations for factor models.

Resources

• GitHub: https://github.com/finmap-org/mcp-server — source, documentation and gpt-actions.yaml
• Hosted endpoint: https://mcp.finmap.org
• npm package: finmap-mcp (run via npx)

If you are integrating the MCP into an agent framework, use the provided MCP tool names or the HTTP endpoints above; the repo contains example client configurations and the OpenAPI schema for automation.