Skip to main content
Glama
dongwoosuk

Outlook MCP

by dongwoosuk
OverviewSchemaRelated ServersScoreDiscussions
Python
Local

📧 Outlook MCP - AEC-Specialized AI Email Intelligence

An MCP (Model Context Protocol) server that transforms Outlook email management with RAG-based semantic search and Vision AI for architectural document analysis - built specifically for Architecture, Engineering & Construction workflows.

Version Status License: MIT Python 3.10+

🎯 Built for AEC Professionals

Unlike generic email tools, this is designed for the unique challenges of architectural practice:

🔍 Semantic Search with RAG

  • ChromaDB vector database for intelligent email indexing

  • Natural language queries: "Find emails about the BIM coordination meeting"

  • Attachment content search: Search inside PDFs, Word docs, Excel files

👁️ Multi-Modal Vision AI

  • Claude Vision API for architectural drawing analysis

  • Gemini Vision API for technical document processing

  • Specialized prompts for floor plans, elevations, sections, details

  • OCR fallback with Tesseract (Korean + English)

🏗️ AEC-Specific Features

  • Bilingual support: Korean + English for international projects

  • Technical drawing analysis: Understands architectural documents

  • RFI/Submittal parsing: Extract key information from construction docs

  • 100% Local processing: Email data never leaves your machine

📋 Features Overview

Feature

Description

Semantic Search

Natural language email queries via RAG

Vision AI

Claude + Gemini for image/PDF analysis

Attachment Parsing

PDF, Word, Excel, Images

Date/Sender Filters

Metadata-based search

Folder Navigation

Access all Outlook folders

Local Processing

Complete privacy, no cloud upload

🛠️ Installation

git clone https://github.com/dongwoosuk/outlook-mcp.git cd outlook-mcp python -m venv .venv .venv\Scripts\activate # Windows pip install -e .

For OCR support:

pip install -e ".[ocr]"

For Vision AI:

pip install anthropic google-generativeai

⚙️ Configuration

Environment Variables (Optional)

# For Vision AI features set ANTHROPIC_API_KEY=your_claude_api_key set GOOGLE_API_KEY=your_gemini_api_key

Claude Desktop Configuration

Add to claude_desktop_config.json:

{ "mcpServers": { "outlook": { "command": "path/to/outlook_mcp/.venv/Scripts/python.exe", "args": ["-m", "outlook_mcp"], "cwd": "path/to/outlook_mcp" } } }

📚 Available Tools

Tool

Description

email_search

Natural language semantic search

email_search_by_sender

Filter by sender

email_search_by_date

Filter by date range

email_search_attachments

Search attachment contents

email_index_status

Check indexing progress

email_index_refresh

Index new emails

email_get_detail

Get full email details

email_list_folders

List Outlook folders

💡 Usage Examples

In Claude Desktop:

"Find emails from the structural engineer about beam calculations" "Search for RFI responses from last week" "Find all emails with PDF attachments about the facade design" "What did the contractor say about the schedule delay?"

Vision AI for Attachments:

"Analyze the floor plan attached in John's email" "What dimensions are shown in this elevation drawing?" "Extract the specification details from the attached PDF"

🏗️ Architecture

outlook_mcp/ ├── outlook_mcp/ │ ├── server.py # Main MCP server │ ├── outlook_reader.py # Win32com Outlook integration │ ├── email_indexer.py # ChromaDB RAG indexing │ ├── attachment_parser.py # PDF/Word/Excel/Vision AI │ ├── config.py # Configuration management │ └── cli.py # Command-line interface └── pyproject.toml

🔧 Technical Stack

Component

Technology

Email Access

win32com (Outlook COM API)

Vector DB

ChromaDB

Embeddings

sentence-transformers (all-MiniLM-L6-v2)

PDF Parsing

PyMuPDF

Word Parsing

python-docx

Excel Parsing

openpyxl

OCR

Tesseract (kor+eng)

Vision AI

Anthropic Claude, Google Gemini

🎯 Use Cases

For Project Architects (PA)

  • Quickly find client feedback across hundreds of emails

  • Search RFI responses and submittal approvals

  • Track consultant coordination threads

For Project Managers (PM)

  • Search contract and schedule discussions

  • Find meeting notes and action items

  • Track change order communications

For Designers

  • Find design review comments

  • Search for reference images and inspiration

  • Locate specification discussions

Future: Office-Wide Deployment

  • Centralized email intelligence for entire teams

  • Shared knowledge base across projects

  • AI-powered email analytics and reporting

⚠️ Requirements

  • Windows with Outlook Desktop app installed and logged in

  • Python 3.10+

  • Outlook must be running for email access

🔒 Privacy & Security

  • 100% local processing: Emails are indexed locally in ChromaDB

  • No cloud upload: Email content never leaves your machine

  • API keys optional: Vision AI features require API keys, but core search works without them

  • Data location: C:\Users\{USERNAME}\Documents\OutlookMCP\

📄 License

MIT License - see LICENSE file.

🙏 Acknowledgments

  • Built on the Model Context Protocol by Anthropic

  • ChromaDB for vector storage

  • sentence-transformers for embeddings

📬 Contact

Dongwoo Suk - Computational Design Specialist

This server cannot be installed

-
security - not tested
A
license - permissive license
-
quality - not tested

How are these scores calculated?

Resources

New MCP Servers

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/dongwoosuk/outlook-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server