mcp-finnhub

mcp-finnhub

MCP server providing comprehensive access to Finnhub financial market data API for AI assistants like Claude Desktop.

Category
Visit Server

README

mcp-finnhub

Python 3.11+ License: MIT MCP

Model Context Protocol (MCP) server providing comprehensive access to Finnhub financial market data API. Built for AI assistants like Claude Desktop to seamlessly access real-time market data, technical indicators, fundamentals, and alternative data.

Features

  • 15 MCP Tools covering 100+ Finnhub API endpoints
  • Real-time market data: Quotes, candles, tick data, BBO
  • Technical analysis: 50+ indicators, patterns, support/resistance
  • Fundamentals: Financials, earnings, dividends, metrics
  • Alternative data: ESG scores, social sentiment, supply chain
  • Multi-asset support: Stocks, crypto, forex, ETFs, bonds
  • Project-based storage with automatic CSV/JSON exports
  • Background job processing for large datasets
  • Smart output handling with token estimation
  • Configurable tools - enable only what you need
  • AI-friendly errors - Structured error responses with guidance
  • Built-in help - Every tool supports operation="help" for discovery

Quick Start

Prerequisites

Installation

Option 1: Install from PyPI (coming soon)

pip install mcp-finnhub

Option 2: Install from source

git clone https://github.com/cfdude/mcp-finnhub.git
cd mcp-finnhub
pip install -e .

Configuration

  1. Create a .env file in your project root:
# Required
FINNHUB_API_KEY=your_api_key_here
FINNHUB_STORAGE_DIR=/path/to/storage

# Optional - Tool configuration
FINNHUB_ENABLE_TECHNICAL_ANALYSIS=true
FINNHUB_ENABLE_STOCK_MARKET_DATA=true
FINNHUB_ENABLE_NEWS_SENTIMENT=true
# ... enable/disable other tools as needed

# Optional - Performance tuning
FINNHUB_RATE_LIMIT_RPM=300  # Requests per minute (30 free, 300 premium)
FINNHUB_REQUEST_TIMEOUT=30  # Request timeout in seconds
FINNHUB_SAFE_TOKEN_LIMIT=75000  # Conservative token limit
  1. Configure Claude Desktop to use mcp-finnhub by adding to claude_desktop_config.json:
{
  "mcpServers": {
    "finnhub": {
      "command": "mcp-finnhub",
      "env": {
        "FINNHUB_API_KEY": "your_api_key_here",
        "FINNHUB_STORAGE_DIR": "/Users/yourname/finnhub-data"
      }
    }
  }
}
  1. Restart Claude Desktop

First Query

Ask Claude:

"Get me the latest quote for AAPL"

Claude will use the finnhub_stock_market_data tool with the get_quote operation to fetch Apple's real-time stock quote.

Available Tools

Core Trading & Analysis (3 tools)

  • finnhub_stock_market_data - Quotes, candles, tick data, BBO, market status
  • finnhub_technical_analysis - 50+ indicators (RSI, MACD, SMA), patterns, signals
  • finnhub_news_sentiment - Company news, market news, sentiment scores

Fundamentals & Analysis (5 tools)

  • finnhub_stock_fundamentals - Financials, earnings, dividends, metrics
  • finnhub_stock_estimates - Revenue/EPS estimates, price targets
  • finnhub_stock_ownership - Insider trades, institutional holdings
  • finnhub_alternative_data - ESG scores, social sentiment, supply chain
  • finnhub_sec_filings - SEC filings, earnings transcripts

Multi-Asset Data (4 tools)

  • finnhub_crypto_data - Crypto exchanges, symbols, profiles, candles
  • finnhub_forex_data - Forex rates, candles, exchanges
  • finnhub_calendar_data - IPO calendar, earnings calendar, economic events
  • finnhub_market_events - Market holidays, analyst upgrades/downgrades

Management Tools (3 tools)

  • finnhub_project_create - Create project workspaces
  • finnhub_project_list - List all projects with statistics
  • finnhub_job_status - Check background job status

Example Workflows

Technical Analysis Workflow

User: "Analyze TSLA using RSI and MACD indicators"

Claude uses:
1. finnhub_technical_analysis → get_indicator (operation=get_indicator, symbol=TSLA, indicator=rsi)
2. finnhub_technical_analysis → get_indicator (operation=get_indicator, symbol=TSLA, indicator=macd)
3. Analyzes both indicators and provides trading insights

Fundamental Research Workflow

User: "Research MSFT - get financials, earnings, and analyst estimates"

Claude uses:
1. finnhub_stock_fundamentals → get_basic_financials (operation=get_basic_financials, symbol=MSFT)
2. finnhub_stock_fundamentals → get_earnings (operation=get_earnings, symbol=MSFT)
3. finnhub_stock_estimates → get_earnings_estimates (operation=get_earnings_estimates, symbol=MSFT)
4. Summarizes financial health and analyst outlook

Multi-Asset Portfolio Workflow

User: "Create a project called 'my-portfolio', get quotes for AAPL, BTC-USD, and EUR/USD"

Claude uses:
1. finnhub_project_create → create (operation=create, project=my-portfolio)
2. finnhub_stock_market_data → get_quote (operation=get_quote, symbol=AAPL, project=my-portfolio)
3. finnhub_crypto_data → get_crypto_profile (operation=get_crypto_profile, symbol=BINANCE:BTCUSDT)
4. finnhub_forex_data → get_forex_rates (operation=get_forex_rates, base=EUR, project=my-portfolio)
5. All data saved to /storage/my-portfolio/ with CSV exports

Tool Configuration

Control which tools are available to reduce context window usage:

# Disable tools you don't need
FINNHUB_ENABLE_CRYPTO=false
FINNHUB_ENABLE_FOREX=false
FINNHUB_ENABLE_ALTERNATIVE_DATA=false

# Core tools (always recommended)
FINNHUB_ENABLE_STOCK_MARKET_DATA=true
FINNHUB_ENABLE_TECHNICAL_ANALYSIS=true
FINNHUB_ENABLE_NEWS_SENTIMENT=true

Rate Limits

  • Free tier: 60 requests/minute
  • Basic tier: 150 requests/minute ($49.99/month)
  • Premium tier: 300 requests/minute
  • Enterprise tier: Custom limits

mcp-finnhub automatically handles rate limiting with exponential backoff and retry logic.

AI Agent Features

Help/Discovery

Every tool supports operation="help" to discover available operations:

finnhub_stock_fundamentals(operation="help")
# Returns all operations with required/optional params and examples

Structured Errors

Errors return actionable JSON instead of stack traces:

{
  "error": "invalid_operation",
  "message": "Operation 'get_data' not found",
  "valid_operations": ["get_basic_financials", "get_dividends", ...],
  "tool": "finnhub_stock_fundamentals"
}

Storage Structure

Projects are stored in subdirectories:

FINNHUB_STORAGE_DIR/
├── my-portfolio/
│   ├── .project.json         # Project metadata
│   ├── candles/              # Historical price data
│   ├── quotes/               # Real-time quotes
│   ├── news/                 # News articles
│   ├── fundamentals/         # Financial data
│   ├── technical/            # Technical indicators
│   └── jobs/                 # Background job results
└── research-tech/
    └── ...

Development

Setup

# Clone repository
git clone https://github.com/cfdude/mcp-finnhub.git
cd mcp-finnhub

# Install with dev dependencies
pip install -e ".[dev]"

# Run tests
pytest --cov=mcp_finnhub --cov-report=term-missing

# Lint and format
ruff check --fix .
ruff format .

Running Locally

# Set environment variables
export FINNHUB_API_KEY=your_api_key
export FINNHUB_STORAGE_DIR=/tmp/finnhub-data

# Run with STDIO transport (default, for Claude Desktop)
mcp-finnhub

# Run with HTTP streaming (recommended for persistent/remote connections)
mcp-finnhub --transport http --port 8000

# Run with SSE transport (deprecated, use HTTP instead)
mcp-finnhub --transport sse --port 8000

# Custom host binding for network access
mcp-finnhub --transport http --host 0.0.0.0 --port 8125

Transport Options

Transport Use Case Command
stdio Claude Desktop, local MCP clients mcp-finnhub (default)
http Remote connections, persistent server, Claude Code mcp-finnhub --transport http --port 8000
sse Legacy SSE clients (deprecated) mcp-finnhub --transport sse --port 8000

Running as a Persistent Service

For HTTP transport, you can run the server as a background service:

# Using nohup (simple)
nohup mcp-finnhub --transport http --port 8000 > finnhub.log 2>&1 &

# Using systemd (Linux)
# See docs/DEPLOYMENT.md for service file examples

Testing with MCP Inspector

# Install MCP inspector
npm install -g @modelcontextprotocol/inspector

# Test the server
mcp-inspector mcp-finnhub

Architecture

  • FastMCP Server: Built on MCP SDK's FastMCP for multi-transport support
  • ServerContext: Dependency injection container managing client and utilities
  • Multi-Transport: STDIO (default), HTTP streaming, and SSE transports
  • Async Job Processing: Background workers for large dataset operations
  • Type Safety: Pydantic models for all API requests/responses

See ARCHITECTURE.md for detailed design documentation.

API Reference

See API.md for complete tool documentation with examples.

Troubleshooting

"Authentication failed" error

  • Verify your FINNHUB_API_KEY is correct
  • Check if your API key has expired
  • Ensure you haven't exceeded rate limits

"Tool not found" error

  • Check that the tool is enabled in your configuration
  • Verify tool names match exactly (case-sensitive)

Storage permission errors

  • Ensure FINNHUB_STORAGE_DIR exists and is writable
  • Check filesystem permissions

Rate limit errors

  • Reduce FINNHUB_RATE_LIMIT_RPM in your configuration
  • Upgrade to premium tier for higher limits

Contributing

Contributions welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Add tests for new functionality
  4. Ensure all tests pass
  5. Submit a pull request

License

MIT License - see LICENSE for details

Acknowledgments

Support


Made with ❤️ for the AI assistant ecosystem

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured