pyp6xer-mcp 1.0.0

MCP server for parsing and analyzing Primavera P6 XER files using PyP6Xer

PyP6Xer MCP Server

AI-powered analysis for Primavera P6 schedules — Parse XER files, analyze critical paths, track progress, and generate insights using Claude Desktop, ChatGPT, or any MCP-compatible AI.

---

Quick Start

Claude Desktop (Recommended)

1. Install via uvx (easiest):

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "pyp6xer": {
      "command": "uvx",
      "args": ["pyp6xer-mcp"]
    }
  }
}

2. Restart Claude Desktop

3. Start analyzing:

"Load the file at /path/to/project.xer and analyze the critical path"

That's it! No installation needed — uvx automatically downloads and runs the latest version.

Alternative: Install locally with pip

pip install pyp6xer-mcp

Then configure Claude Desktop:

{
  "mcpServers": {
    "pyp6xer": {
      "command": "pyp6xer-mcp"
    }
  }
}

Alternative: Run via Docker

docker compose up

Then point Claude Desktop to the HTTP endpoint:

{
  "mcpServers": {
    "pyp6xer": {
      "url": "http://localhost:5000/mcp"
    }
  }
}

---

What Can It Do?

📊 Schedule Analysis

"What's on the critical path?"
"Show me activities with negative float"
"Run a DCMA 14-point schedule health check"
"Which activities are slipping?"

📈 Progress & Performance

"What's the current progress summary?"
"Calculate earned value metrics (SPI, CPI)"
"Show me work package status by responsible manager"
"Compare baseline to actual performance"

🔍 Resource Management

"List all resources and their assignments"
"Show me resource utilization over time"
"Which resources are over-allocated?"

📁 Data Export

"Export the critical path to CSV"
"Generate an executive summary report"
"Export schedule dashboard for Excel"

22 tools • 5 workflow prompts • 3 data resources — See TOOLS.md for complete reference.

---

Installation Methods

Method 1: uvx (Recommended)

No installation needed! Just add to claude_desktop_config.json:

{
  "mcpServers": {
    "pyp6xer": {
      "command": "uvx",
      "args": ["pyp6xer-mcp"]
    }
  }
}

Method 2: pip

pip install pyp6xer-mcp

Method 3: Docker

# Clone repo
git clone <repo-url>
cd pyp6xer_mcp

# One-command setup
docker compose up

Method 4: From Source (Developers)

# Clone repo
git clone <repo-url>
cd pyp6xer_mcp

# Install with uv
uv pip install -e .

# Run tests
pytest

---

Example Workflows

🎯 Critical Path Analysis

You: "Load the file at /Users/me/Desktop/Q1_Project.xer and show me the critical path"

Claude:

• Loads the XER file
• Identifies critical activities
• Shows duration, dates, and relationships
• Highlights longest path through the network

🏥 Schedule Health Check

You: "Run a schedule health check on the loaded project"

Claude:

• Checks for DCMA 14-point compliance
• Identifies missing logic, constraints, lags
• Flags activities with high/negative float
• Provides recommendations

📊 Progress Dashboard

You: "Generate a progress summary and export to CSV"

Claude:

• Calculates percent complete by project, WBS, activity
• Shows planned vs actual dates
• Exports formatted CSV for Excel pivot tables

🔄 Baseline Comparison

You: "Compare baseline to actual and identify variances"

Claude:

• Compares planned vs actual start/finish dates
• Calculates schedule variance
• Identifies slipping activities
• Recommends corrective actions

💼 Executive Summary

You: "Create an executive summary for the steering committee"

Claude:

• Overall project status (on track, at risk, behind)
• Key milestones and dates
• Critical path activities
• Resource concerns
• Recommendations

---

Configuration

Claude Desktop Config Location

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Linux: ~/.config/Claude/claude_desktop_config.json

Environment Variables

Create a .env file for local development:

# Optional: API key for HTTP mode
API_KEY=your_secret_key_here

# Optional: Port (defaults to 5000)
PORT=5000

---

Running as HTTP Server

For remote access or web UI integration:

# Without auth (development only!)
python -m pyp6xer_mcp.cli streamable-http

# With auth (production)
API_KEY=your_secret_key python -m pyp6xer_mcp.cli streamable-http

Live deployment: https://pyp6xer-mcp.fly.dev/mcp

Upload Files via HTTP

curl -X POST -H "Authorization: Bearer <api_key>" \
  -F "file=@project.xer" \
  https://pyp6xer-mcp.fly.dev/upload

Response:

{
  "cache_key": "abc123",
  "projects": [...],
  "activity_count": 250,
  "message": "XER loaded. Use cache_key 'abc123' with MCP tools."
}

---

Troubleshooting

"Server not found" in Claude Desktop

1. Check that claude_desktop_config.json is in the correct location
2. Ensure JSON is valid (no trailing commas)
3. Restart Claude Desktop completely (Quit, not just close window)
4. Check Claude Desktop logs:
   • macOS: ~/Library/Logs/Claude/mcp*.log
   • Windows: %APPDATA%\Claude\logs\mcp*.log

"ModuleNotFoundError: No module named 'pyp6xer_mcp'"

Using uvx? No action needed — it auto-installs.

Using pip? Run:

pip install pyp6xer-mcp

"Failed to parse XER file"

• Ensure file is valid P6 XER format (not XML, not .xls)
• Try opening in P6 to verify file integrity
• Check file size (very large files may timeout)

Permission Denied

# macOS/Linux: Grant file access
chmod +r /path/to/file.xer

# Or move file to accessible location
mv /restricted/path/file.xer ~/Desktop/

---

Architecture

Package Structure

pyp6xer_mcp/
├── __init__.py           # Package entry point
├── server.py             # FastMCP server instance
├── cli.py                # CLI entry point
├── models/               # Pydantic input models
│   ├── common.py         # ResponseFormat, ExportType
│   ├── file_ops.py       # Load, write, export models
│   ├── activities.py     # Activity models
│   ├── analysis.py       # Analysis models
│   └── resources.py      # Resource models
├── core/                 # Core utilities
│   ├── cache.py          # XerCache (in-memory)
│   ├── formatters.py     # Format functions
│   └── helpers.py        # Helper functions
├── tools/                # 22 MCP tools
│   ├── file_ops.py       # load, write, export, clear
│   ├── projects.py       # list_projects
│   ├── activities.py     # list, get, search, update
│   ├── analysis.py       # critical_path, float, quality
│   ├── resources.py      # list_resources, utilization
│   ├── progress.py       # progress, earned_value
│   └── calendars.py      # list_calendars
├── prompts/              # 5 workflow prompts
│   └── analysis.py       # Preset workflows
├── resources/            # 3 MCP resources
│   └── schedule.py       # Direct data access
└── http/                 # HTTP layer
    ├── middleware.py     # Auth middleware
    └── routes.py         # Health, upload endpoints

Tool Categories

Category	Count	Tools
File Operations	4	load_file, write_file, clear_cache, export_csv
Project/Activity	5	list_projects, list_activities, get_activity, search_activities, update_activity
Schedule Analysis	6	critical_path, float_analysis, schedule_quality, relationship_analysis, slipping_activities, schedule_health_check
Resource Management	2	list_resources, resource_utilization
Progress/Performance	4	progress_summary, earned_value, work_package_summary, wbs_analysis
Structure	1	list_calendars

See TOOLS.md for complete documentation.

---

Development

Run Tests

# All tests (27 tests)
pytest

# With coverage
pytest --cov=pyp6xer_mcp

# Specific test file
pytest tests/test_modular_structure.py -v

Linting & Formatting

black --line-length 100 .
ruff check .
mypy .

Build & Publish

# Build package
python -m build

# Publish to PyPI
python -m twine upload dist/*

---

Requirements

• Python: 3.10+
• Dependencies:
  • mcp>=1.0.0 - MCP protocol
  • pydantic>=2.0.0 - Input validation
  • PyP6XER>=0.1.0 - XER parsing
  • httpx>=0.24.0 - HTTP client
  • uvicorn>=0.23.0 - ASGI server

---

Need Help?

💼 Consulting Services

Need custom schedule analysis, integration, or P6 automation?

Contact: [Your consulting page/email]

Services:

• Custom MCP tool development
• P6 schedule optimization
• Integration with project management systems
• Training and workshops

Minimum engagement: $800

🐛 Bug Reports & Feature Requests

Open an issue on GitHub Issues (coming soon)

📚 Documentation

• TOOLS.md - Complete tool reference
• CLAUDE.md - Developer documentation
• Model Context Protocol Docs

---

License

MIT License - see LICENSE for details.

Free for commercial and personal use. No attribution required.

---

Acknowledgments

• PyP6Xer by Hassan Emam - XER parsing library
• Model Context Protocol by Anthropic - Protocol specification
• FastMCP - Python MCP framework