jmabry/mcp-tarrifs

3.1

If you are the rightful owner of mcp-tarrifs and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

A comprehensive MCP server implementation for accessing and analyzing USITC tariff data.

Tools

2

Resources

0

Prompts

0

USITC Tariff Data MCP Server

Query 11 years of US tariff data through AI assistants using MCP (Model Context Protocol).

Quick Start

Build database:

python scripts/mcp_server_launcher.py --build-only

Configure your MCP client (Claude, Cursor, etc.):

{
  "name": "usitc-tariffs",
  "command": "uv", 
  "args": ["run", "python", "scripts/mcp_server_launcher.py"],
  "env": {}
}

Query tariff data:
- "What's the tariff rate for cars from Germany?"
- "Compare steel tariff rates over the past 3 years"
- "Show me HTS codes for electronics"

Available Tools

get_tariff_rates - Find tariff rates by HTS code or product search
compare_tariff_rates - Compare rates across multiple years
query_database - Direct SQL queries on tariff database

Data Coverage

Years: 2015-2024 (11 years)
Records: 142K+ tariff entries
Source: USITC Annual Tariff Database
Database: DuckDB (local file)

Development

# Run tests
pytest

# Start development server
python scripts/mcp_server_launcher.py

# Build database with specific years
python src/tariffs_db/db_build.py --years 5

Client Configuration Examples

Claude Desktop

{
  "mcpServers": {
    "usitc-tariffs": {
      "command": "uv",
      "args": ["run", "python", "scripts/mcp_server_launcher.py"],
      "cwd": "/path/to/mcp-tarrifs"
    }
  }
}

Cursor

{
  "name": "usitc-tariffs",
  "command": "uv",
  "args": ["run", "python", "scripts/mcp_server_launcher.py"]
}

Requirements

Python 3.8+
uv (recommended) or pip
~2GB disk space for database

License

MIT