MCP Configuration Examples

Example MCP (Model Context Protocol) manifests live in examples/mcp_configurations/. Use them to scope which tools are exposed to Claude or other MCP clients. If you installed from PyPI, copy the manifests from the repo or create your own.

Quick Setup

For guided setup, run:

pip install "fmp-data[mcp]"
fmp-mcp setup

See Claude Desktop Setup for full setup instructions.

Available Configurations

Default (no manifest)

Uses the full tool set. Do not set FMP_MCP_MANIFEST.

minimal_manifest.py

The smallest useful configuration with just essential tools: - Company profiles and quotes - Basic market data - Cryptocurrency support

Use case: Testing, development, or lightweight deployments

trading_manifest.py

Optimized for active trading and technical analysis: - Real-time quotes and market movers - Technical indicators (SMA, RSI, MACD, etc.) - Pre-/post-market data - Price targets and news

Use case: Day trading, swing trading, technical analysis

research_manifest.py

Comprehensive tools for fundamental analysis: - Complete financial statements - Key metrics and ratios - Analyst recommendations - Institutional holdings - Economic indicators

Use case: Investment research, due diligence, long-term investing

crypto_manifest.py

Specialized for cryptocurrency markets: - Crypto quotes and historical data - Intraday crypto prices - Crypto news - Technical indicators for crypto

Use case: Cryptocurrency trading and analysis

Usage

Quick Setup with CLI (Recommended)

# Setup wizard for Claude Desktop
fmp-mcp setup

# Check server status
fmp-mcp status

# Test server connection
fmp-mcp test

# List available tools
fmp-mcp list

# List tools for specific client
fmp-mcp list --client market

Manual Configuration

Using with Python module execution

export FMP_API_KEY=your_api_key_here  # pragma: allowlist secret
export FMP_MCP_MANIFEST=examples/mcp_configurations/trading_manifest.py
python -m fmp_data.mcp

Using with MCP CLI

FMP_API_KEY=your_api_key mcp dev python -c "
from fmp_data.mcp.server import create_app
app = create_app("examples/mcp_configurations/research_manifest.py")
app.run()
"

Using in your own code

from fmp_data.mcp.server import create_app

# Load a specific configuration
app = create_app(tools="examples/mcp_configurations/minimal_manifest.py")
app.run()

Creating Custom Configurations

You can create your own manifest by:

1. Manual creation - Copy one of these examples and modify:

   # my_custom_manifest.py
   TOOLS = [
       "company.profile",
       "company.quote",
       # Add your desired tools here
   ]
   

2. Using the CLI tool - Generate a manifest with specific tools:

   # Generate manifest with specific tools
   fmp-mcp generate my_manifest.py --tools company.profile company.quote
   
   # Generate manifest without default tools
   fmp-mcp generate my_manifest.py --no-defaults --tools company.quote market.gainers
   

3. Discovering available tools - List all available tools:

   # List all tools in table format
   fmp-mcp list
   
   # List as tree structure
   fmp-mcp list --format tree
   
   # Filter by client module
   fmp-mcp list --client technical
   
   # Export as JSON
   fmp-mcp list --format json > tools.json
   

Tip: set FMP_MCP_TOOL_NAME_STYLE=spec to expose fully qualified tool names (client.key) and avoid naming collisions when multiple tools share a key.

Validation

Validate your manifest file before using:

fmp-mcp validate my_manifest.py

Tips

• Start with minimal_manifest.py and add tools as needed
• Group related tools together with comments
• Test your configuration with a simple query before deployment
• Consider API rate limits when selecting tools
• Use research_manifest.py for comprehensive analysis
• Use trading_manifest.py for real-time market monitoring