h1b-mcp

An MCP server that lets Claude answer H1B visa sponsorship questions using free public DOL data.

Ask things like:

• "Does Google sponsor H1Bs?"

• "What's Amazon's H1B approval rate and top roles?"

• "Which companies in Austin sponsor H1Bs for Data Engineers?"

Tools

Setup

1. Clone and install

git clone https://github.com/<your-username>/h1b-mcp.git
cd h1b-mcp
python3 -m venv venv
source venv/bin/activate      # Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Build the database (one-time, ~2-3 min)

python loader.py

This downloads the FY2024 Q2 DOL H1B LCA data (~50MB) and loads it into a local SQLite database.

3. Verify the database

python -c "import sqlite3; db = sqlite3.connect('h1b.db'); print(db.execute('SELECT COUNT(*) FROM lca_applications').fetchone())"

Should print a non-zero count (e.g., (600000,)).

4. Connect to Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "h1b": {
      "command": "/absolute/path/to/h1b-mcp/venv/bin/python",
      "args": ["/absolute/path/to/h1b-mcp/server.py"]
    }
  }
}

Important: Use the absolute path to the venv Python, not bare python. To find it:

cd h1b-mcp && source venv/bin/activate && which python

5. Restart Claude Desktop and start asking

"Which companies in New York sponsor H1Bs for software engineers?"

Data Source

Public DOL LCA Disclosure Data (FY2024 Q2). Source: https://www.dol.gov/agencies/eta/foreign-labor/performance

No API keys required. Data is a quarterly snapshot — re-run loader.py to refresh.

Running Tests

source venv/bin/activate
pytest tests/ -v