Resume Screening MCP Service

A service for screening resumes using Large Language Models. This service evaluates resumes against specific requirements using semantic understanding rather than simple keyword matching.

Installation

Using UV (recommended)

uv pip install .

Using pip

pip install .

Configuration

The application uses environment variables for configuration. Create a .env file in the root directory with the following variables:

OPENAI_API_KEY=your_api_key_here
OPENAI_MODEL=gpt-4o-mini

Usage

Command Line Interface

# Process resumes in a specific directory
resume-screening --resume-dir=./my_resumes --report-dir=./reports

# Or using environment variables
RESUME_DIR=./my_resumes REPORT_DIR=./reports resume-screening

As MCP Service

from fastmcp import FastMCPClient

client = FastMCPClient()
response = await client.call("resume_screening", {
    "prompt": "筛选精通Python且有金融机构从业经验的简历",
    "resume_dir": "./resumes",
    "report_dir": "./reports"
})
print(response)

Directory Structure

• resume_screening_mcp/ - Main package directory
  • __init__.py - Package initialization
  • core.py - Core resume processing functionality
  • cli.py - Command line interface
  • server.py - MCP server implementation

Development

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

# Run tests
pytest

License

MIT