kellykohlleffel/fivetran-mcp-toolkit

Fivetran MCP Toolkit

Manage your Fivetran data pipelines directly from Claude Desktop using the Model Context Protocol.

What is This?

The Fivetran MCP Toolkit enables Claude to interact with your Fivetran account through natural language. Manage data connections, monitor sync status, invite users, and more—all through conversation with Claude Desktop.

Example:

You: List all my Fivetran connections

Claude: I found 12 connections in your account:
• postgres_production (PostgreSQL) - Syncing every 6 hours
• salesforce_main (Salesforce) - Syncing every 24 hours
• stripe_payments (Stripe) - Paused
[...]

Features

• 25 Core Tools for essential Fivetran operations
• Natural Language Interface - Just ask Claude what you need
• Full Connection Management - Create, delete, pause, resume, and sync connections
• User Management - Invite users, modify roles, and manage access
• Schema Operations - View and reload connector schemas
• Transformation Control - List and run dbt transformations
• Destination Management - Manage data warehouses and groups
• Real-time Monitoring - Check sync status and health

Installation

Prerequisites

• Python 3.10 or higher
• Claude Desktop
• Fivetran account with API access

Quick Start

1. Clone the repository:

git clone https://github.com/yourusername/fivetran-mcp-toolkit.git
cd fivetran-mcp-toolkit

2. Install dependencies:

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

3. Get Fivetran API credentials:

• Log in to Fivetran
• Click your username in the left navigation
• Click API Key
• Click Generate API Key
• Save both the API Key and API Secret securely

4. Configure Claude Desktop:

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "fivetran": {
      "command": "python3",
      "args": [
        "/absolute/path/to/fivetran-mcp-toolkit/src/mcp_server.py"
      ],
      "env": {
        "FIVETRAN_API_KEY": "your_api_key_here",
        "FIVETRAN_API_SECRET": "your_api_secret_here"
      }
    }
  }
}

5. Restart Claude Desktop

6. Test it:

Ask Claude: "List my Fivetran connections"

Detailed setup: See

Available Tools (25)

Connection Management (8 tools)

User Management (4 tools)

Destination Management (3 tools)

Connector Metadata (4 tools)

Transformations (3 tools)

Groups (3 tools)

Full API Reference: See

Usage Examples

Check Connection Status

You: What's the status of my PostgreSQL connection?

Claude: [Calls get_sync_status]
Your PostgreSQL connection is healthy:
• Status: Connected
• Last sync: 2 hours ago (successful)
• Next sync: In 4 hours
• Sync frequency: Every 6 hours

Pause During Maintenance

You: We're doing database maintenance. Pause the production connections.

Claude: [Calls pause_connection for each]
Paused 3 connections:
  • postgres_production
  • mysql_orders
  • mongodb_events

Invite Team Member

You: Invite sarah@company.com as an Account Analyst

Claude: [Calls invite_user]
Invitation sent to sarah@company.com
Role: Account Analyst
They'll receive setup instructions via email.

Monitor Issues

You: Are any connections failing?

Claude: [Analyzes connections]
Found 1 issue:
• mysql_orders - Connection timeout (last failed 3 hours ago)
  Recommendation: Check database connectivity

More Examples: See

Documentation

Roadmap

v1.0 (Current)

• 25 core tools covering essential operations
• Connection, user, and destination management
• Schema operations and transformations
• Comprehensive documentation

Additional Tools Available

The Fivetran API supports many more operations that you can add as needed:

• Bulk operations (pause/resume/sync multiple)
• Usage analytics and performance reports
• Certificate and security management
• Multi-account support
• Advanced transformation management

See for the complete list and how to extend the toolkit.

Contributing

We welcome contributions! Please see our for details on:

• Reporting bugs
• Suggesting features
• Submitting pull requests
• Development setup
• Coding standards

License

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

Security

• Credentials are never logged or stored
• All API communication uses HTTPS
• Runs locally on your machine
• No data sent to third parties

Troubleshooting

Common Issues:

Detailed Help: See

Acknowledgments

• Built on the Model Context Protocol by Anthropic
• Powered by the Fivetran REST API
• Inspired by the MCP community and Infinite Lambda's MCP Fivetran Package Server

Support

• Report Bug
• Request Feature
• Discussions

Show Your Support

If you find this toolkit useful, please consider:

• Starring the repository
• Reporting bugs
• Suggesting features
• Contributing improvements
• Sharing with others

Made with care for the data engineering community

Note: This is an unofficial community project and is not affiliated with or endorsed by Fivetran Inc.