Your AI assistant can now access your actual database schema in real-time. No more copy-pasting table structures or explaining relationships every conversation.

This guide shows you how to set up Guepard's MCP server with Cursor IDE in under 10 minutes. Once configured, your AI will know your database structure, generate accurate queries, and provide schema-aware suggestions automatically.

What you'll achieve: Schema-aware AI assistance that understands your specific database without manual context switching.

Time investment: 10 minutes setup, hours saved weekly.

‍

What is MCP? (Model Context Protocol)

MCP is a protocol that lets AI assistants connect directly to your development tools and data sources. Instead of your AI living in isolation, it can access your actual database schemas, APIs, and documentation in real-time.

‍

‍

Want to learn more? Check the official MCP documentation

‍

Why This Matters for Database Work

Schema Awareness: Your AI knows your exact table names, column types, foreign keys, and indexes without you explaining them.

Real-time Updates: When you add tables or modify schemas, your AI assistant automatically sees the changes.

Context Preservation: Complex relationships and constraints are understood immediately, not explained repeatedly.

Security: MCP provides read-only access to metadata only. Your actual data stays protected.

‍

MCP vs. Traditional AI Chat

‍

MCP transforms your AI from a general assistant to a specialized teammate that understands your specific development environment.

‍

Guepard MCP: Database Environments for AI

Guepard MCP connects your AI assistant to your database deployments, branches, and snapshots. Your AI can see your actual schema, create test environments, and manage database operations directly.

‍

‍

Core Features

• Schema Access: AI knows your exact table structures, relationships, and constraints
• Environment Management: Create, start, stop database deployments from AI chat
• Branch Operations: Switch between database branches, create test environments
• Snapshot Control: Take snapshots, restore from specific points in time
• Real-time Updates: AI sees changes as you modify your database structure
• Multi-Database: Support for PostgreSQL and other database engines

‍

What Your AI Can Do

• Generate queries that work on your actual schema
• Create isolated test environments for experimentation
• Manage database branches like Git branches
• Take snapshots before risky operations
• Clone production data for safe testing
• Monitor deployment status and compute resources

‍

Database Operations Available

• test_connection - Verify MCP server connectivity
• list_deployments - View all your database environments
• create_deployment - Launch new database instances
• start_compute / stop_compute - Manage database resources
• create_snapshot - Save current database state
• list_branches - See all database branches
• create_branch - Fork database for testing
• checkout_branch - Switch active database branch

Documentation: docs for complete reference and advanced features.

‍

Requirements Checklist

Before we dive in, let's make sure you have everything needed. This takes about 5 minutes to verify, and prevents the "why isn't this working" frustration later.

‍

Essential Tools

Cursor IDE

• Download from cursor.sh if you don't have it
• Why Cursor? Built-in AI chat with MCP support - no extensions needed

Quick check: Open Cursor and press Cmd/Ctrl + i to verify AI chat works

‍

‍

Node.js (v18 or higher)

• Check your version: node --version
• Need to install? Grab it from nodejs.org

‍

‍

Docker Desktop

• Download from docker.com/products/docker-desktop
• Must be running (not just installed)

Quick verification: docker --version should return version info

‍

‍

Get Your Guepard Access Token

1. Sign up or login at guepard.run

‍

‍

2. Go to Access Tokens in the sidebar

‍

‍

3. Click "Generate Token"

‍

‍

4. Copy the token (you'll paste this in Step 4)

‍

‍

• Save it: This token won't be shown again

‍

Step-by-Step Setup Guide (5 Minutes to Database-Aware AI)

You've got the tools, you've got the token. Time to connect your AI to your actual database schema.

Step 1: Open Cursor

• Launch Cursor IDE

Step 2: Open Settings

• Click on the Settings Icon (top right)

‍

‍

Step 3: Find MCP Settings

• Click on "MCP" section

‍

‍

Step 4: Add Guepard MCP Configuration

• Click "Add Custom MCP"
• A file mcp.json should open, enter this configuration:

‍

{
"mcpServers": {
  "Guepard": {
    "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "access_token=<GUEPARD_ACCESS_TOKEN>",
        "mghassen/guepard-mcp-server:1.4.0"
      ]
    }
  }
}

‍

• Replace <GUEPARD_ACCESS_TOKEN> with the token you copied in Step 4 of requirements

‍

‍

• Save and close

‍

Step 5: Start Docker

• Open Docker Desktop
• Ensure it shows "Running" status

‍

Step 6: Refresh MCP Connection

• Back in Cursor MCP settings, refresh MCP Server
• Wait for green status indicator next to "Guepard"

‍

‍

• If red: Double-check your Access Token and that Docker is running

‍

Step 7: Test Your Setup

• Open a new AI chat in Cursor (Cmd/Ctrl + i)
• Type: "give me my databases"
• Your AI should respond with your deployed databases in Guepard

‍

‍

You're Done

Your AI assistant now has direct access to your Guepard database environments. No more explaining schemas, no more context switching.

Quick verification checklist:

• ✅ Green MCP status in settings
• ✅ AI responds to Guepard-specific commands
• ✅ No authentication errors in Cursor logs

Troubleshooting:

• Red MCP status? Verify API key and restart Docker
• AI doesn't see deployments? Check your Guepard account has active environments
• Connection timeout? Your network might be blocking the MCP server

Next: See this database-aware AI in action with real development scenarios.

Commands list: https://www.guepard.run/mcp

‍

Start Your Database-Aware Development Journey

You've seen the setup and the potential. Integrating your AI with your database schema isn't just a technical novelty, it's a fundamental shift in how you interact with your data layer.

Next Steps:

• Set up your environment: If you haven't already, follow the steps in this guide. It's a one-time, 10-minute commitment.
• Connect your first database: Experience the "aha" moment when the AI instantly understands your tables without any manual explanation.
• Experience true schema-aware AI: Ask it a complex question about your data. You'll never want to go back to copy-pasting context again.

The Developer Reality:

• What changes: Your daily relationship with database development becomes a conversation, not a chore. You focus on business logic, not boilerplate SQL.
• Time investment: A brief setup saves you hours of manual query writing, schema checking, and debugging every single week.
• Team impact: Onboarding new developers becomes easier. Junior team members can write safe, effective queries. Senior developers can focus on architecture instead of repetitive tasks.

Ready to Transform Your Database Development?

• Start your free account: guepard.run/signup
• Follow this guide: Bookmark this page for a smooth setup.
• Join the community: Share your experience and see how others are leveraging database-aware AI in their workflows.

‍