Miles Free AI Library
← Back to resources
GuideInstagramSep 29, 2026

Connect Claude to FMP for Stock and Portfolio Research

Configure the Financial Modeling Prep connector and use ready-to-copy prompts to screen stocks, investigate insider buying, run portfolio DCF valuations, and create recurring market briefs.

Get the full MD

Download the complete Claude and FMP research guide

Free
Download MD

Connect Claude to FMP for Stock and Portfolio Research

Set up an FMP data connector, then use three copy ready prompts to screen stocks, examine insider activity, and value a portfolio.

What you will achieve

You will configure Claude to access FMP market data and run structured financial research using current prices, company statements, insider transactions, analyst ratings, and other available datasets.

Who this is for

Investors and analysts who want to use Claude as a data assisted research tool.

Difficulty: Intermediate

Short tutorial

Set up an FMP data connector, then use three copy ready prompts to screen stocks, examine insider activity, and value a portfolio.

Guide: Connect Claude to FMP for Market and Portfolio Research

Time: 20 to 30 minutes

This tutorial shows you how to connect Claude to Financial Modeling Prep (FMP), then use the connection for stock screening, transaction research, and discounted cash flow analysis. Follow the setup first, test the data connection, and then adapt the supplied prompts to your portfolio or research criteria.

FMP access does not make Claude a licensed financial adviser. Treat its output as research, verify important figures against the underlying records, and make investment decisions based on your own circumstances.

What the connection enables

According to the source material, the FMP connector can give Claude access to datasets including:

  • Financial statements
  • Current market prices
  • Insider transactions
  • Analyst ratings
  • Stock and cryptocurrency data

The exact data available may depend on your FMP account, plan, connector configuration, and the relevant endpoint.

1. Create and configure your FMP account

  1. Create an account with FMP.
  2. Open the FMP dashboard.
  3. Locate and copy your API key.
  4. Store the key securely. Do not place it in public prompts, screenshots, shared chats, or documents.

The source states that a free account is enough to begin, with an upgrade available if greater usage or coverage is needed. Check your account limits before attempting broad market screens.

2. Add FMP as a Claude connector

In Claude:

  1. Open Settings.
  2. Select Connectors.
  3. Choose Add a Custom Connector.
  4. Enter the FMP connector link.
  5. Save the connector.
  6. When prompted during the first request, provide your FMP API key.

The supplied material does not include the exact connector URL, so obtain it from the FMP connector instructions rather than guessing it.

3. Verify the connection before doing analysis

Start with a narrow request that makes it easy to inspect the result:

TEXT
Using the connected FMP data, retrieve the latest available price, reported revenue, operating cash flow, and the reporting dates for [TICKER].

For every figure:
1. State the currency and reporting period.
2. Identify whether it is annual, quarterly, or current market data.
3. Include the retrieval time or latest available date.
4. If a field is unavailable, say so instead of estimating it.

Check that Claude actually used the connector and that the reporting periods align. A current price combined with an older financial statement is not necessarily wrong, but those dates must be clear.

4. Find stocks that declined while fundamentals improved

This screen looks for companies whose prices fell by more than 15 percent during the current month even though revenue and cash flow continued to grow.

TEXT
Using FMP data, screen the available stock universe for companies whose share price is down more than 15% during the current calendar month but whose revenue and cash flow are still growing.

For each match, provide:
1. Ticker and company name.
2. Percentage price decline and the exact measurement dates.
3. Revenue growth rate and the periods compared.
4. Cash flow growth rate, naming the cash flow metric used and the periods compared.
5. Recent financial or market evidence that may explain the decline.
6. An assessment of whether the decline appears justified by the available numbers.
7. Key risks or missing data that could change the conclusion.

Rank the results from strongest to weakest fundamental profile. Do not infer unavailable values. Clearly distinguish reported facts from interpretation.

How to review the screen

Confirm that every result satisfies all three conditions:

  • Price declined by more than 15 percent in the specified period.
  • Revenue grew across clearly identified comparable periods.
  • The selected cash flow measure grew across comparable periods.

Pay particular attention to the definition of cash flow. Operating cash flow and free cash flow are different measures, so the output should identify which one it used.

5. Examine clustered insider purchases

The second workflow searches for companies where at least three insiders purchased shares recently, then checks for disclosed congressional purchases in the same securities.

TEXT
Using the available FMP datasets, find stocks where three or more distinct company insiders reported open market share purchases during the last 30 calendar days.

For each stock, show:
1. Company name and ticker.
2. Each insider's name and role, if available.
3. Transaction date, filing date, number of shares, price, and transaction value.
4. Whether the transactions appear to be separate purchases by distinct insiders.
5. Whether the available data shows a disclosed purchase of the same stock by a member of Congress.
6. The congressional transaction date, disclosure date, reported value or range, and purchaser, if available.
7. Any missing fields, reporting delays, or ambiguity in transaction type.

Exclude grants, option exercises, awards, and automatic transactions unless they also represent a clearly identified open market purchase. Rank results by the number of distinct insider buyers, then by total reported insider purchase value. Do not treat these transactions alone as a recommendation to buy.

This distinction matters because an insider filing can represent several transaction types. Review the transaction classification instead of assuming every increase in ownership was a voluntary market purchase.

The source claims that congressional transaction data can be checked through this workflow. If that dataset is not exposed by your connector or account, Claude should report that limitation rather than present an unsupported match.

6. Run discounted cash flow analysis across a portfolio

Prepare a simple portfolio list before using the final prompt:

TEXT
AAPL: 20 shares
MSFT: 10 shares
NVDA: 5 shares

Avoid including account numbers, personal identifiers, or other unnecessary private information.

Then run this prompt:

TEXT
Using FMP financial statements and the latest available market prices, run a discounted cash flow analysis for every stock in the portfolio below.

Portfolio:
[PASTE TICKERS AND SHARE COUNTS HERE]

For each company:
1. State the valuation date and current price used.
2. Show the historical financial inputs used.
3. State the cash flow measure used as the valuation base.
4. Explain the revenue growth, margin, reinvestment, discount rate, and terminal value assumptions.
5. Calculate estimated intrinsic value per share.
6. Calculate the percentage discount or premium to the current price.
7. Provide bear, base, and bull cases.
8. Identify the assumptions that have the greatest effect on the result.
9. Flag stale, missing, inconsistent, or unusual data.

Rank the holdings from most undervalued to most overvalued under the base case. Then identify possible candidates to add to, hold, or trim, explaining the valuation evidence and uncertainty behind each classification. Present this as research support, not personalized financial advice.

Required quality checks

Before relying on the ranking, confirm that the response includes:

  • A valuation date for every stock
  • Explicit assumptions rather than unexplained estimates
  • Consistent treatment of cash, debt, and share count
  • More than one valuation scenario
  • Warnings where the model is highly sensitive
  • Clear separation between calculated value and current price

A discounted cash flow result can change substantially when growth, discount rate, margins, or terminal assumptions change. The scenario table is therefore as important as the base case ranking.

7. Create a reusable market brief

If your Claude environment supports scheduled tasks, the same connector can support a recurring portfolio update. Use this template:

TEXT
Using the connected FMP data, prepare a concise market and portfolio brief for the holdings below:

[PASTE TICKERS HERE]

Include:
1. Latest available price and percentage change for each holding.
2. Material changes in reported financial data.
3. New insider transactions.
4. New analyst rating information available through the connector.
5. The largest portfolio price movements.
6. Items that need further investigation.

Show the date and time of the data. Separate confirmed data from interpretation. If no new information is available for a category, state that clearly. Do not invent explanations for price movements.

Final operating checklist

  • Confirm Claude used FMP rather than relying only on general knowledge.
  • Require dates, reporting periods, currencies, and metric definitions.
  • Reduce the scope if the free account cannot screen the full market.
  • Ask Claude to expose missing data instead of filling gaps with estimates.
  • Verify important filings and transactions before acting.
  • Treat screens and valuation rankings as starting points for deeper research, not automatic trading instructions.
Continue exploring

What to watch next