Personal Productivity MCP Server

A multi-domain Model Context Protocol (MCP) server providing Claude with tools to manage different aspects of personal productivity.

Overview

This MCP server provides tools across three domains:

1. Career (Phase 1 - Current): Job search automation, application tracking, resume analysis

2. Fitness (Phase 3 - Planned): Health tracking, workout management, progress analysis

3. Family (Phase 4 - Planned): Kids billing, expense tracking, Math Academy monitoring

Current Features (Phase 1)

Career Domain Tools

Read Tools:

• career_search_greenhouse - Search jobs on Greenhouse job boards

• career_search_lever - Search jobs on Lever job boards

• career_get_applications - Get tracked applications with optional status filter

Write Tools:

• career_track_application - Track a new job application

• career_update_application - Update application status and notes

Analysis Tools:

• career_analyze_pipeline - Get statistics about your application pipeline

Quick Start

Prerequisites

• Docker and Docker Compose

• Claude Desktop or compatible MCP client

Setup

1. Clone the repository:

2. Copy the environment template:

3. Build and run with Docker:

Connecting to Claude Desktop

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

Or for local development without Docker:

Usage Examples

Searching for Jobs

Claude will use the career_search_greenhouse and career_search_lever tools to find available positions.

Tracking Applications

Viewing Your Pipeline

Project Structure

Development

Local Development Setup

1. Create a virtual environment:

2. Install dependencies:

3. Install Playwright browsers:

4. Run the server:

Running Tests

Code Formatting

Configuration

Environment variables (see .env.example):

• MCP_SERVER_PORT - Server port (default: 3000)

• LOG_LEVEL - Logging level (default: INFO)

• DATABASE_PATH - Path to SQLite database

• CACHE_TTL_JOBS - Job listing cache TTL in seconds (default: 3600)

• PLAYWRIGHT_HEADLESS - Run Playwright in headless mode (default: true)

Roadmap

• Phase 1: Basic career domain with Greenhouse/Lever scrapers

• Phase 2: Complete career domain with resume analysis

• Phase 3: Fitness domain with health tracking

• Phase 4: Family domain with billing automation

• Phase 5: Polish and monitoring

Architecture Decisions

Why Docker?

• Avoids local Python environment issues (especially with Playwright)

• Consistent deployment across different machines

• Easy to run in home lab environment

Why SQLite?

• Simple file-based storage, no separate database server needed

• Perfect for single-user personal productivity use case

• Easy backups (just copy the database file)

Why MCP?

• Native integration with Claude

• Structured tool interface with validation

• Async-first design for scraping and API calls

Contributing

This is a personal project, but feel free to fork and adapt for your own needs!

License

MIT