TMorville/wyndle

🌿 Wyndle - Your Slack Conversation Assistant

"I must be helpful! It is very important that I be helpful!"

Transform your Slack data into an intelligent personal assistant using AI and high-performance DuckDB storage.

✨ Features

🧠 AI-Powered Intelligence

• Smart Summaries: "Nothing urgent from Emil! He wishes you good holiday" instead of raw message dumps
• Priority Management: Automatically identifies what needs your attention with urgency levels
• Relationship Analysis: Understand conversation dynamics and outcomes across all contexts

⚡ High-Performance Data Pipeline

• DuckDB Storage: Columnar database for sub-millisecond query performance
• Continuous Sync: Background loader with intelligent rate limiting (respects Slack's 45 calls/minute)
• Human-Readable: All data uses real names instead of cryptic IDs

🔗 MCP Integration

• Model Context Protocol: Works seamlessly with Raycast, OpenAI, and other AI tools
• Natural Queries: Ask "What needs my attention?" and get actionable insights
• Configurable: Smart bot filtering and customizable response styles

🚀 Quick Start

Prerequisites

• Python 3.10+
• UV package manager (recommended) or pip
• Slack workspace with bot token

Installation

# Clone the repository
git clone https://github.com/yourusername/wyndle.git
cd wyndle

# Install dependencies
uv sync

# Configure your setup
cp config/config.example.yaml config/config.yaml
cp .env.example .env

# Set up the database (IMPORTANT: Required before first run)
cp data/conversations.example.duckdb data/conversations.duckdb

⚠️ Important: You must copy the example database before running your first --dataloader command. This creates the proper database schema for storing your Slack data.

Configuration

1. Slack Bot Setup: Create a Slack app with these scopes:

   channels:history, groups:history, im:history, mpim:history
   users:read, channels:read
   

2. Environment Variables: Add to .env:

   SLACK_USER_KEY=xoxp-your-user-token-here
   

3. Channel Selection: Edit config/config.yaml:

   slack:
     channels:
       - general
       - engineering
       - data-team
     ignored_bots:
       - slackbot
       - github
       - jira
   

Usage

# One-time data load
uv run wyndle-pipeline --dataloader

# Start continuous background sync
uv run wyndle-loader

# Launch MCP server for AI integration
uv run wyndle-server

# View database statistics
uv run wyndle-pipeline --database-stats

🎯 Raycast Integration

Setup Instructions

1. Install MCP Extension in Raycast

   • Open Raycast
   • Go to Extensions → Store
   • Search for "MCP" and install the MCP extension

2. Install MCP Server

   • Complete the installation steps above first
   • Make sure you have your .env file configured with API keys

3. Add Wyndle to Raycast

   • Open Raycast MCP settings
   • Click "Add MCP Server"
   • Command: <your-path-to-repo>/scripts/run-wyndle.sh
   • Name: Wyndle
   • Save the configuration

Usage

Once configured, you can interact with Wyndle directly in Raycast:

@Wyndle please look at my squad channel, make a list of action points for me and add them as todo's with @todoist

@Wyndle anything to follow up on today?

@Wyndle summarize my latest interaction with Emil B

@Wyndle show me what's happening in the data team channel this week

Wyndle will provide intelligent, context-aware responses and can integrate with other Raycast extensions like Todoist for task management.

💬 AI Integration Examples

With Raycast

"What needs my attention?"
→ 🔥 **Urgent**: Sarah needs feedback on design mockups
→ 📅 **This Week**: Confirm availability for bi-weekly data meetings

"Summarize my interaction with John"
→ ✅ **Project complete** - John thanked you for the pipeline work and signed off

With OpenAI/Claude

The MCP server provides these intelligent tools:

• content_get_user_interactions() - Relationship summaries
• productivity_list_followups() - Action items and priorities
• content_get_channel_activity() - Team/project intelligence

🏗️ Architecture

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   Slack API     │───▶│   DuckDB        │───▶│   MCP Server    │
│                 │    │   (Columnar)    │    │   (FastMCP)     │
│ • Rate Limited  │    │ • Human Names   │    │ • AI Assistant  │
│ • Continuous    │    │ • Sub-ms Query  │    │ • Smart Filter  │
│ • Smart Sync    │    │ • 70% Smaller   │    │ • Raycast Ready │
└─────────────────┘    └─────────────────┘    └─────────────────┘

Key Components

• src/ingest/ - Continuous data pipeline with rate limiting
• src/data/ - DuckDB storage layer with name resolution
• src/server.py - Personal assistant MCP server
• src/cli.py - Command-line interface
• src/slack_client/ - Slack API integration
• src/analysis/ - Writing style and organizational analysis

🧪 Development

Running Tests

# Run minimal test suite
uv run pytest tests/

# Type checking
uv run mypy src/

# Code quality
uv run ruff check .
uv run ruff format .

🤝 Contributing

We welcome contributions! Please see our for details.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes with tests
4. Run the test suite (uv run pytest)
5. Submit a pull request

📝 License

This project is licensed under the MIT License - see the file for details.

🙏 Acknowledgments

• Built with FastMCP for Model Context Protocol integration
• Powered by DuckDB for high-performance analytics
• Inspired by the need for intelligent Slack data management

📊 Performance

• Query Speed: Sub-millisecond response times with DuckDB columnar storage
• Memory Usage: ~300MB constant footprint with efficient compression
• Storage: 70% smaller than equivalent SQLite databases
• Rate Limiting: Respects Slack's API limits with intelligent backoff