If you use Claude for investment research or financial analysis, there is a good chance you have run into its biggest limitation: Claude does not have access to live market data. It can reason about stocks, explain financial concepts, and build models — but it cannot tell you what Apple is trading at right now.
That changes with MCP.
What Is MCP and Why Does It Matter for Stock Data?
MCP stands for Model Context Protocol. It is an open standard that lets AI assistants like Claude connect to external data sources in real time. Think of it as giving Claude a phone line to call out and get information it would not otherwise have.
MarketXLS has built an MCP server that exposes over 500 financial functions directly to Claude. That means you can ask Claude to fetch a stock price, pull an options chain, look up a company's balance sheet, or compare dividend yields across a watchlist — and it will actually go get that data live, right in the conversation.
The server URL is:
https://qm-mcp.marketxls.com/mcp
It uses OAuth 2.0 authentication tied to your MarketXLS account, so your API usage and permissions carry over seamlessly.
What Can You Actually Do With This?
Here is the practical part. Once connected, you can ask Claude things like:
- "What is the current price of AAPL?"
- "Show me the options chain for TSLA expiring this Friday with bid and ask prices."
- "Compare the P/E ratios of MSFT, GOOG, and META."
- "Pull the last 30 days of closing prices for SPY and calculate the standard deviation."
- "What is the dividend yield for the top 10 holdings in my portfolio?"
Claude will use the MarketXLS MCP tools to fetch the data, then analyze it, format it, or build on it however you ask. You are not copying and pasting from a terminal. You are having a conversation with an AI that has direct access to market data.
The Four MCP Tools
The MarketXLS MCP server gives Claude four tools:
-
search_functions — Search through 500+ available financial functions by keyword. Ask Claude to find functions related to "options," "dividends," "earnings," or anything else.
-
get_function_details — Get the full documentation for a specific function, including parameters, return values, and examples.
-
call_function — Execute a single function with the parameters you specify. This is how Claude actually fetches a stock price or pulls financial data.
-
batch_call_functions — Run multiple function calls at once. Useful when you want to compare several tickers or pull different data points for the same stock in a single request.
How to Set It Up in Claude Desktop
Setting this up takes about two minutes.
Step 1: Open Your Claude Desktop Config
On Mac, open the file at:
~/Library/Application Support/Claude/claude_desktop_config.json
On Windows, it is at:
%APPDATA%\Claude\claude_desktop_config.json
Step 2: Add the MarketXLS MCP Server
Add the following to your config file:
{
"mcpServers": {
"marketxls": {
"url": "https://qm-mcp.marketxls.com/mcp"
}
}
}
Step 3: Restart Claude Desktop
Close and reopen Claude Desktop. It will detect the new MCP server and prompt you to authenticate with your MarketXLS account via OAuth.
That is it. You are connected.
A Real Example: Fetching Stock Prices in a Conversation
Here is what a typical interaction looks like once MCP is set up:
You: "What is NVDA trading at right now, and how does that compare to its 52-week high?"
Claude: Uses call_function to fetch the current price and 52-week high for NVDA, then responds with the numbers and calculates how far the stock is from its high as a percentage.
No browser tabs. No spreadsheet. No API keys to manage. Just a question and an answer backed by real data.
Why This Is Useful for Investors and Analysts
The value here is not just convenience. It is the combination of live data and reasoning in the same place.
When you ask Claude to "screen for undervalued tech stocks with a P/E below 20 and positive free cash flow," it can actually pull the data for each candidate, filter the results, and explain its reasoning — all in one conversation. Before MCP, you would need to pull that data separately, paste it in, and hope Claude interpreted your formatting correctly.
With MCP, the data comes in structured. Claude knows exactly what each number represents because it called the function itself.
Common Use Cases
- Portfolio review: Ask Claude to pull current prices and key metrics for all your holdings, then flag anything that has moved significantly.
- Options analysis: Fetch a full options chain with bid, ask, Greeks, and open interest, then ask Claude to identify interesting spreads or unusual activity.
- Earnings prep: Pull revenue, EPS, and guidance history before an earnings call, and have Claude summarize the trends.
- Screening: Describe what you are looking for in plain English, and let Claude search for the right functions, pull the data, and filter the results.
- Comparing investments: Ask Claude to compare two or more stocks across any combination of fundamental, technical, or options metrics.
More Than 500 Functions Available
The MarketXLS function library covers a wide range of market data, including:
- Real-time and delayed stock quotes
- Historical price data
- Options chains with full Greeks, bid/ask, and open interest
- Fundamental data: income statements, balance sheets, cash flow
- Dividend history and yields
- Technical indicators
- ETF holdings and sector breakdowns
- Analyst estimates and ratings
- Earnings dates and results
- Institutional ownership data
You can ask Claude to search for available functions on any topic using the search_functions tool. If you are not sure what data is available, just ask — Claude will look it up for you.
How Is This Different From ChatGPT or Other AI Tools?
Most AI tools with market data access rely on plugins or pre-built integrations that expose a limited set of endpoints. The MCP approach is different because it gives Claude access to the full MarketXLS function library — the same 500+ functions available in the Excel add-in — through a standardized protocol.
This means you are not limited to whatever a plugin developer decided to include. If MarketXLS has a function for it, Claude can call it.
MCP is also an open standard, not a proprietary plugin format. As the protocol matures, the same MarketXLS MCP server will work with any AI assistant that supports MCP.
Getting Started
To use the MarketXLS MCP server, you need:
- A MarketXLS subscription (any plan that includes API access)
- Claude Desktop (or any MCP-compatible client)
- Two minutes to add the server URL to your config
If you are already a MarketXLS user, you likely have everything you need. Visit the MarketXLS MCP documentation page for the full list of available functions, or check out the MCP setup guide for detailed instructions.
If you are new to MarketXLS, you can start with a free trial and connect to the MCP server right away.
The Bottom Line
Fetching stock pricing data into Claude using MCP is straightforward. You add one URL to your config, authenticate with your MarketXLS account, and start asking questions backed by real market data. No API keys to juggle, no code to write, no data to format.
For anyone who already uses Claude for financial thinking, this closes the gap between reasoning and data. You no longer have to bring the numbers to the AI — the AI goes and gets them.
