Meeting Scheduler MCP
Enables natural language meeting scheduling by searching IMAP emails for meeting requests, finding available calendar slots from YAML configuration, and creating email drafts with proper threading to maintain personal conversation context.
README
Meeting Scheduler MCP
Natural Language Meeting Scheduling with Personal Touch
The Problem with Scheduling Tools
Traditional scheduling tools like Calendly and Cal.com feel transactional for important contacts:
- Impersonal: Automated links lack human touch
- Manual Hell: Calendar tab-switching between email and scheduling apps
- Context Loss: No integration with email conversations
Our Goal: Natural language scheduling that feels personal and integrates seamlessly with your email workflow.
What We're Building
A Mistral Le Chat powered meeting scheduler that combines:
- Email Search: Find meeting requests in your inbox
- Calendar Management: YAML-based scheduling with holidays support
- Natural Language Interface: Chat-based scheduling workflow
- Email Threading: Maintain conversation context with proper email headers
Architecture Overview
sequenceDiagram
participant U as User
participant L as Mistral Le Chat
participant M as MCP Server
participant I as IMAP Server
participant Y as calendar.yaml
U->>L: "Find a meeting request from Lisa"
activate L
L->>M: search_emails(query="Lisa")
activate M
M->>I: IMAP SEARCH
activate I
I-->>M: Email results
deactivate I
M-->>L: Email
deactivate M
L->>M: get_free_slots()
activate M
M->>Y: Read free slots
activate Y
Y-->>M: Available times
deactivate Y
M-->>L: Free slots
deactivate M
L-->>U: "Lisa wants to meet. You're free Tue 2pm or Thu 10am"
U->>L: "Suggest Thursday, keep it casual"
L-->>U: Draft preview
U->>L: "Send it"
L->>M: save_draft_and_block_slot(...)
activate M
M->>Y: Write blocked slot
M->>I: IMAP APPEND to INBOX.Drafts
M-->>L: Confirmation
deactivate M
L-->>U: "Draft saved, slot blocked"
deactivate L
Tech Stack
- Backend: Python 3.13+ with FastMCP
- Email Integration: IMAP with TLS/SSL support
- Calendar Storage: YAML file-based system
- Transport: HTTP/SSE for Mistral connector
- Data Models: Pydantic v2 for validation
Features
1. Email Integration
- IMAP Search: Find emails by criteria (sender, subject, etc.)
- Email Threading: Maintain conversation context with Message-ID, In-Reply-To, References
- Draft Management: Save email drafts with proper threading headers
- TLS/SSL Support: Secure connections with configurable SSL options
2. Calendar Management
- YAML Configuration: Simple, human-readable calendar format
- Weekly Availability: Define working hours and available slots
- Holiday Support: Automatic holiday detection (country-specific)
- Time Zone Awareness: Full timezone support with ZoneInfo
- Slot Finding: Intelligent free slot detection
3. Natural Language Workflow
- Find Requests: "Find meeting requests from Lisa"
- Get Availability: "What times am I free this week?"
- Suggest Times: "Suggest Thursday morning"
- Confirm & Block: "Confirm the meeting and block the slot"
Getting Started
Prerequisites
- Python 3.13+
- IMAP-enabled email account
- Mistral Le Chat integration
Installation
# Clone the repository
uv add git+https://github.com/seb-schulz/meeting-scheduler-mcp.git
# Install dependencies
uv sync
# Set up environment variables
cp .env.example .env
# Edit .env with your IMAP credentials
Configuration
Create a .env file with your IMAP settings:
# Required settings
IMAP_HOST=imap.example.com # IMAP server hostname
IMAP_USER=your@email.com # IMAP username/email
IMAP_PASSWORD=secure_password # IMAP password
IMAP_FROM=your@email.com # From address for sent emails
# TLS/SSL Configuration (optional with sensible defaults)
IMAP_PORT=993 # IMAP port
IMAP_USE_SSL=true # Use SSL/TLS
IMAP_USE_STARTTLS=false # Use STARTTLS
IMAP_VERIFY_SSL=true # Verify SSL certificates
Calendar Configuration
Edit calendar.yaml to define your availability:
schedule:
timezone: Europe/Berlin
slot_duration: 30
holidays: DE
weekly:
- days:
- mon
- tue
- wed
- thu
- fri
slots:
- start: 09:00:00
end: "12:00:00"
- start: "13:00:00"
end: "17:00:00"
blocked: []
Running the Server
# Start the FastMCP server with Python module syntax
uv run python -m meeting_scheduler_mcp
# Server will be available at http://0.0.0.0:8000
API Tools
1. search_emails
Search emails with full metadata and threading support:
emails = search_emails(
mailbox="INBOX",
criteria="FROM lisa@example.com"
)
# Returns: List of emails with id, subject, from, to, date, message_id, in_reply_to, references, body
2. get_free_slots
Get available time slots from your calendar:
free_slots = get_free_slots()
# Returns: List of up to 50 available slots with date, start, end times, and timezone
3. save_draft_and_block_slot
Complete workflow: block calendar slot and save confirmation email draft:
result = save_draft_and_block_slot(
datetime="2025-12-15T14:00:00+01:00",
duration=30,
reason="Meeting with Lisa",
subject="Meeting Confirmed",
body="Let's meet on Thursday at 2pm",
to="lisa@example.com",
in_reply_to="<original-message-id@example.com>"
)
# Returns: {"success": true} or {"error": "message", "success": false}
Testing
Unit Tests
Run tests with:
uv run pytest
Integration Tests with Docker Compose
For IMAP integration testing, use Docker Compose with Greenmail:
# Start Greenmail IMAP server
cd .devcontainer
docker-compose up -d greenmail
# Seed test emails
uv run python scripts/seed_test_emails.py
# Run tests with test configuration
IMAP_CONFIG=.env.test uv run pytest tests/test_integration.py
Test Coverage
Run tests with coverage:
uv run pytest --cov=meeting_scheduler_mcp --cov-report=term-missing
Contributing
Contributions welcome! Please follow:
- Code Style: Ruff linting rules
- Type Safety: Full Pydantic v2 typing
- Testing: Comprehensive pytest coverage
- Documentation: Keep README updated
License
MIT License - See LICENSE for details.
Learn More
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。