findata-mcp 0.3.0

Financial data MCP server — stocks, fundamentals, FRED economics, SEC filings, crypto via x402 micropayments

FinData MCP

The first comprehensive financial data MCP server. Five tools covering real-time stocks, company fundamentals, US economic indicators (FRED), SEC filings, and cryptocurrency — all in one server.

Pay per call via x402 micropayments ($0.01/call) or register a free API key (1,000 calls/day). No signup required to start.

---

Tools

Tool	Description	Cache
stock_quote(ticker)	Real-time (15-min delayed) price, volume, and change %	1 min
company_fundamentals(ticker)	Revenue, P/E, market cap, sector, beta, dividend yield	1 hr
economic_indicator(series_id)	800,000+ FRED series: GDP, CPI, fed funds rate, yield curves	6 hr
sec_filing(ticker_or_cik, form_type)	Full text of 10-K, 10-Q, 8-K from SEC EDGAR	24 hr
crypto_price(coin_id)	Price, market cap, 24h volume, 7-day sparkline via CoinGecko	1 min

---

Installation

Claude Desktop (Smithery)

smithery install findata-mcp

uvx (any MCP client)

uvx findata-mcp

pip

pip install findata-mcp
findata-mcp

Docker

docker run -p 8080:8080 \
  -e FINDATA_API_KEY=fd_live_your_key \
  cortexlabs/findata-mcp

---

Configuration

All environment variables are optional. The server works out of the box via x402 micropayments.

Variable	Description	Default
FINDATA_API_KEY	API key for rate-limited access	—
FRED_API_KEY	FRED API key for higher rate limits	—
X402_WALLET_ADDRESS	Wallet address to receive x402 payments	placeholder
X402_NETWORK	Payment network (eip155:84532 = Base Sepolia)	Base Sepolia
X402_VERIFY	Verify x402 payments on-chain (true/false)	false
FINDATA_DB_PATH	Path for API key SQLite database	./findata.db

---

Pricing

Tier	Price	Limit
x402 pay-per-call	$0.01 / call	Unlimited
Free API key	$0	1,000 calls/day
Pro API key	$9/month	10,000 calls/day

Register at findata-mcp.io/keys.

---

Usage Examples

Stock quote

# MCP tool call
stock_quote(ticker="AAPL")

# Returns:
{
  "ticker": "AAPL",
  "price": 213.49,
  "change": 1.23,
  "change_pct": 0.58,
  "volume": 54321000,
  "market_cap": 3280000000000,
  "currency": "USD",
  "timestamp": "2026-03-01T15:30:00Z"
}

Economic indicator (FRED)

economic_indicator(series_id="FEDFUNDS")

# Returns:
{
  "series_id": "FEDFUNDS",
  "title": "Federal Funds Effective Rate",
  "units": "Percent",
  "frequency": "Monthly",
  "latest_value": 4.33,
  "latest_date": "2026-02-01",
  "observations": [...]
}

SEC filing

sec_filing(ticker_or_cik="AAPL", form_type="10-K")

# Returns:
{
  "ticker": "AAPL",
  "cik": "320193",
  "form_type": "10-K",
  "filed_at": "2025-10-31",
  "period": "2025-09-27",
  "filing_url": "https://www.sec.gov/...",
  "text_excerpt": "..."
}

---

Architecture

FinData MCP has two components: a thin client (what you install) and a hosted backend (what runs on Railway).

Thin Client (findata_mcp/) — shipped via PyPI

The pip install / uvx package is a lightweight MCP stdio server that proxies requests to the hosted backend. It handles x402 micropayment signing automatically. No API keys, no local data providers needed.

findata_mcp/
├── __init__.py        # Entry point (findata-mcp CLI)
├── server.py          # FastMCP thin server (5 tool stubs)
└── client.py          # HTTP client with x402 auto-payment

Requires: EVM_PRIVATE_KEY env var — a wallet funded with USDC on Base mainnet ($0.01/call).

Hosted Backend (backend/) — deployed on Railway

The full implementation with data providers (Yahoo Finance, FRED, SEC EDGAR, CoinGecko), caching, metering, and x402 payment verification. Not included in the PyPI package.

backend/
├── app.py             # FastAPI HTTP server (x402 payments + metering)
├── cache.py           # In-memory TTL cache
├── metering.py        # Call logging + usage tracking
├── server.py          # Legacy full local MCP server (not used in production)
└── tools/
    ├── stock_quote.py
    ├── company_fundamentals.py
    ├── economic_indicator.py
    ├── sec_filing.py
    └── crypto_price.py

Live at: https://findata-mcp-production-1cd3.up.railway.app

---

Development

git clone https://github.com/sapph1re/findata-mcp
cd findata-mcp

# Install thin client in dev mode
pip install -e .

# Run thin client (connects to hosted backend)
findata-mcp

# Run backend locally
pip install -r requirements.txt
cd backend && uvicorn app:app --host 0.0.0.0 --port 8080

# Run tests
python -m pytest test_thin_client.py -v

---

License

MIT — see LICENSE

Built by Cortex Labs