JobSyncd

AI-powered, intelligent automation for job application tracking using LLMs and MCPs.

Features

• 📧 Gmail Integration: Fetches recent job application emails
• 🤖 AI Email Parsing: Uses LLM to extract structured data from emails
• 📊 Notion Database: Automatically creates entries in your Notion job tracker
• 🔄 Daily Sync: Prevents duplicate entries with smart caching
• 🔁 Smart Retry Logic: Automatically handles OpenRouter rate limits with backoff
• 🎯 Application ID Tracking: Updates existing entries when status changes
• 📈 Weekly Reports: AI-generated summaries with statistics and insights

Architecture

Project Structure

JobSyncd/
├── mcp_servers/           # MCP servers for external services
│   ├── gmail_server.py    # Gmail MCP server
│   ├── notion_server.py   # Notion MCP server
│   └── weekly_report_server.py # Weekly report MCP server
├── workflows/             # LangGraph workflows
│   ├── job_sync_workflow.py # Main job sync workflow
│   └── weekly_report_workflow.py # Weekly report workflow
├── agent/                 # Main automation agents
│   ├── gmail_client.py    # Gmail API client
│   ├── credentials.json   # Gmail OAuth credentials (download from Google Cloud Console)
│   ├── token.json         # OAuth token (auto-generated)
│   ├── notion_utils.py    # Notion database operations
│   ├── main.py            # Daily sync orchestrator (MCP + LangGraph)
│   └── weekly_report.py   # Weekly summary generator
├── pyproject.toml         # Project configuration
├── requirements.txt       # Python dependencies
├── uv.lock                # Locked dependencies
├── README.md              # Project overview and quick start
└── SETUP_ENV.md           # Complete setup guide (includes weekly reports)

Data Flow

• ✅ MCP Architecture: Modular, reusable services
• ✅ LangGraph Workflow: Intelligent email processing with built-in deduplication
• ✅ Single LLM Call: Processes all emails together for better context
• ✅ Smart Deduplication: LLM naturally understands email relationships
• ✅ Extensible: Easy to add new services and workflows

Setup

See SETUP_ENV.md for complete setup instructions including Gmail API, Notion configuration, and environment variables.

Usage

Test Notion Connection

uv run agent/notion_utils.py

Run Daily Sync

uv run agent/main.py

This will:

1. Fetch recent emails (last 10 emails by default)
2. Process all emails together with AI for intelligent deduplication
3. Extract job application info with built-in duplicate detection
4. Create entries in your Notion database
5. Handle status updates automatically

On first run: A browser will open for Gmail OAuth authentication. Grant permissions and the agent will save a token.json for future use.

Prerequisites:

• You need to download credentials.json from Google Cloud Console and place it in the agent/ folder
• See SETUP_ENV.md for detailed setup instructions

Key improvements:

• Smart Deduplication: LLM processes all emails together, naturally identifying related emails
• Status Progression: Automatically handles Applied → Assessment → Interview → Offer flows
• No Duplicates: Same company emails are merged into single applications

Generate Weekly Report

Note: Requires Weekly Reports database setup (see SETUP_ENV.md).

uv run agent/weekly_report.py

This will:

1. Fetch application data from the last 7 days using MCP + LangGraph
2. Generate AI-powered summary with intelligent insights
3. Create a formatted report in your Notion database
4. Handle rate limits and errors gracefully

Custom time range:

uv run agent/weekly_report.py 14  # Last 14 days
uv run agent/weekly_report.py 30  # Last 30 days (monthly report)

Key improvements:

• MCP Architecture: Modular, reusable weekly report services
• LangGraph Workflow: Intelligent data processing and summary generation
• Smart Insights: AI analyzes patterns and provides actionable recommendations
• Flexible Timeframes: Support for custom date ranges

Scheduling

GitHub Actions (Recommended - Cloud-based)

Automate everything with GitHub Actions - no need to keep your computer running!

Features:

• ☁️ Runs in the cloud (free for public repos)
• 📅 Daily sync + weekly reports
• 🔄 Auto-refreshes OAuth tokens
• 💾 Persists cache between runs
• 📧 Email notifications on failures

Setup (GitHub Actions):

• Run once locally: uv run agent/main.py (creates agent/token.json)
• Add GitHub Secrets with raw JSON content (Settings → Secrets → Actions):
  • GMAIL_CREDENTIALS = contents of agent/credentials.json
  • GMAIL_TOKEN = contents of agent/token.json
  • NOTION_TOKEN, NOTION_DATABASE_ID, NOTION_WEEKLY_REPORTS_DB_ID, OPENROUTER_KEY
• Ensure workflow permissions: Settings → Actions → General → Read and write
• Manually run each workflow once to verify

See “GitHub Actions Setup” in SETUP_ENV.md for details.

Schedules:

• Daily Sync: Every day at 9:00 AM UTC
• Weekly Report: Every Monday at 10:00 AM UTC

---

Local Scheduling (Alternative)

If you prefer running locally on your own machine:

Daily Sync

Windows (Task Scheduler):

schtasks /create /tn "JobSyncd-Daily" /tr "uv run D:\Projects\JobSyncd\agent\main.py" /sc daily /st 09:00

Linux/Mac (Cron):

crontab -e
# Add: 0 9 * * * cd /path/to/JobSyncd && uv run agent/main.py

Weekly Reports

Windows (Task Scheduler):

schtasks /create /tn "JobSyncd-Weekly" /tr "uv run D:\Projects\JobSyncd\agent\weekly_report.py" /sc weekly /d MON /st 10:00

Linux/Mac (Cron):

crontab -e
# Add: 0 10 * * 1 cd /path/to/JobSyncd && uv run agent/weekly_report.py

Using Python schedule library

import schedule
import time

schedule.every().day.at("09:00").do(daily_sync)

while True:
    schedule.run_pending()
    time.sleep(3600)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License. See the LICENSE file for more details.