Vybe MCP - Second Memory System

An MCP (Model Context Protocol) server that provides a "second memory" system for AI assistants. Store, retrieve, and share summaries through Discord.

Features

• Summarize & Store: Automatically summarize text and store it in a SQLite database
• Memory Resources: Access stored memories as MCP resources in new chats/tabs
• Discord Integration: Send summaries to Discord channels as .txt files
• Search: Search through stored memories by content
• Docker Support: Run everything in containers

Quick Start

1. Install Dependencies

bun install

2. Configure Discord (Optional)

Copy .env.example to .env and fill in your Discord bot credentials:

cp .env.example .env

Edit .env:

DISCORD_BOT_TOKEN=your_bot_token_here
DISCORD_CHANNEL_ID=your_channel_id_here

How to get Discord credentials:

1. Go to https://discord.com/developers/applications
2. Create a new application
3. Go to "Bot" section → Create a bot
4. Copy the bot token
5. Enable "Message Content Intent" in Bot settings
6. Get your channel ID (right-click channel → Copy ID)

3. Run the Server

# Development mode
bun run dev

# Production mode
bun run start

# With MCP Inspector (for testing)
bun run inspector

Docker Setup

Build and Run

# Build image
bun run docker:build
# or
docker build -t vybe-mcp .

# Run container (automatically loads .env file)
bun run docker:run
# or
docker run -i --rm --env-file .env vybe-mcp

# Or use Docker Compose (automatically loads .env file)
docker compose up --build

With MCP Inspector

# Using npm script (automatically loads .env file)
bun run docker:inspector

# Or manually
npx @modelcontextprotocol/inspector docker run -i --rm --env-file .env vybe-mcp

MCP Tools

summarize_and_store

Summarizes text and stores it in the database.

Parameters:

• text (required): The text to summarize
• tags (optional): Array of tags for categorization
• maxLength (optional): Maximum summary length (50-1000, default: 200)
• metadata (optional): Additional metadata object

Example:

{
  "text": "Long conversation about project planning...",
  "tags": ["project", "planning"],
  "maxLength": 300
}

send_to_discord

Sends a summary to Discord channel as a .txt file.

Parameters:

• memoryId (optional): ID of stored memory to send
• text (optional): Text to send directly
• includeMetadata (optional): Include metadata in file (default: true)

Example:

{
  "memoryId": 1
}

search_memories

Search through stored memories.

Parameters:

• query (required): Search query
• limit (optional): Max results (1-50, default: 10)

Example:

{
  "query": "project planning",
  "limit": 5
}

MCP Resources

memory://{id}

Retrieve a specific memory by ID.

Example: memory://1

memory://all

Get all stored memories (up to 50 most recent).

Example: memory://all

memory://stats

Get memory statistics and recent memories.

Example: memory://stats

Usage in Cursor/Claude

Add to your MCP configuration (~/.cursor/mcp.json):

{
  "mcpServers": {
    "vybe-mcp": {
      "command": "bun",
      "args": ["run", "/path/to/vybe-mcp/src/index.ts"],
      "env": {
        "DISCORD_BOT_TOKEN": "your_token",
        "DISCORD_CHANNEL_ID": "your_channel_id"
      }
    }
  }
}

Or with Docker:

{
  "mcpServers": {
    "vybe-mcp": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-v", "/path/to/data:/app/data",
        "--env-file", "/path/to/.env",
        "vybe-mcp"
      ]
    }
  }
}

Workflow Example

1. Store a summary:

   Use summarize_and_store to save: "Discussed project timeline with team"

2. In a new chat/tab:

   Access memory://all to see all stored memories

3. Send to Discord:

   Use send_to_discord with memoryId: 1

4. Search:

   Use search_memories with query: "project"

Database

The database is stored in data/memories.db (SQLite). The schema includes:

• id: Unique memory ID
• original_text: Full original text
• summary: Generated summary
• created_at: Creation timestamp
• updated_at: Last update timestamp
• tags: JSON array of tags
• metadata: JSON object with additional data

Project Structure

vybe-mcp/
├── src/
│   ├── index.ts          # Main MCP server
│   ├── db.ts             # Database operations
│   ├── summarizer.ts     # Text summarization
│   └── discord.ts         # Discord integration
├── data/                  # Database storage (created automatically)
├── Dockerfile
├── docker-compose.yml
├── package.json
└── README.md

Development

# Watch mode
bun run dev

# Test with inspector
bun run inspector

# Build
bun run build

License

MIT