Integration Guide - SnackBase

This guide shows you how to integrate the SnackBase MCP server with various AI assistants and development tools.

Claude Code Integration

Claude Code is a CLI tool that provides AI assistance directly in your terminal. It has native MCP support.

Setup

Install Claude Code (if you haven’t already):

npm install -g claude-code

Configure MCP Server:

Edit ~/.claude/settings.json (or %APPDATA%\claude\settings.json on Windows):

{
  "mcpServers": {
    "snackbase": {
      "command": "snackbase-mcp",
      "args": [],
      "env": {
        "SNACKBASE_URL": "https://your-snackbase-instance.com",
        "SNACKBASE_API_KEY": "your-api-key"
      }
    }
  }
}

Restart Claude Code to load the MCP server.

Usage Examples

Once configured, you can interact with your SnackBase data naturally:

# List all users
claude "List all users in my account"

# Create a record
claude "Create a new post in the posts collection with title 'Hello World'"

# Query with filters
claude "Show me all posts created in the last 7 days"

# Get statistics
claude "What are my dashboard statistics?"

Common Workflows

Data Exploration

You: "Show me the structure of my 'products' collection"
Claude: [Uses snackbase_collections tool to get collection details]

Quick CRUD Operations

You: "Create a new user with email [email protected]"
Claude: [Uses snackbase_users tool to create user]

Data Analysis

You: "What are the top 10 most viewed articles?"
Claude: [Uses snackbase_records tool with sort and limit]

Cursor IDE Integration

Cursor is an AI-powered code editor with MCP support.

Setup

Install Cursor from cursor.sh
Configure MCP Server:

Edit ~/.cursor/settings.json (or %APPDATA%\cursor\settings.json on Windows):

{
  "mcpServers": {
    "snackbase": {
      "command": "snackbase-mcp",
      "args": [],
      "env": {
        "SNACKBASE_URL": "https://your-snackbase-instance.com",
        "SNACKBASE_API_KEY": "your-api-key"
      }
    }
  }
}

Restart Cursor to load the MCP server.

Usage in Cursor

Use Cmd/Ctrl + K to open the AI chat
The AI will have access to your SnackBase data
Ask questions about your data naturally

Continue.dev Integration

Continue.dev is a VS Code extension for AI code assistance.

Setup

Install Continue.dev from the VS Code marketplace
Configure MCP Server:

Edit your VS Code settings.json or Continue config:

{
  "continue.mcpServers": {
    "snackbase": {
      "command": "snackbase-mcp",
      "args": [],
      "env": {
        "SNACKBASE_URL": "https://your-snackbase-instance.com",
        "SNACKBASE_API_KEY": "your-api-key"
      }
    }
  }
}

Usage in VS Code

Use Cmd/Ctrl + Shift + L to open Continue
Ask questions about your SnackBase data
Get code suggestions based on your schema

Cline (Formerly Claude Dev) Integration

Cline is an autonomous coding agent for VS Code.

Setup

Install Cline from the VS Code marketplace
Configure MCP Server in Cline’s settings:

{
  "cline.mcpServers": {
    "snackbase": {
      "command": "snackbase-mcp",
      "args": [],
      "env": {
        "SNACKBASE_URL": "https://your-snackbase-instance.com",
        "SNACKBASE_API_KEY": "your-api-key"
      }
    }
  }
}

Custom Integration

If you want to build your own MCP client, you can use the MCP SDK directly:

TypeScript Example

import { Client } from '@modelcontextprotocol/sdk/client/index.js';
import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';

async function main() {
  // Create transport to the MCP server
  const transport = new StdioClientTransport({
    command: 'snackbase-mcp',
    args: [],
    env: {
      SNACKBASE_URL: 'https://your-snackbase-instance.com',
      SNACKBASE_API_KEY: 'your-api-key',
    },
  });

  // Create and connect client
  const client = new Client({
    name: 'my-mcp-client',
    version: '1.0.0',
  });

  await client.connect(transport);

  // List available tools
  const tools = await client.listTools();
  console.log('Available tools:', tools);

  // Call a tool
  const result = await client.callTool({
    name: 'snackbase_collections',
    arguments: {
      action: 'list',
    },
  });

  console.log('Collections:', result);
}

Python Example

import asyncio
import subprocess
import json

async def call_mcp_tool(tool_name, arguments):
    """Call an MCP tool via stdio"""
    process = subprocess.Popen(
        ['snackbase-mcp'],
        stdin=subprocess.PIPE,
        stdout=subprocess.PIPE,
        stderr=subprocess.PIPE,
        text=True,
        env={
            'SNACKBASE_URL': 'https://your-snackbase-instance.com',
            'SNACKBASE_API_KEY': 'your-api-key',
        }
    )

    # Send request
    request = {
        "jsonrpc": "2.0",
        "id": 1,
        "method": "tools/call",
        "params": {
            "name": tool_name,
            "arguments": arguments
        }
    }

    process.stdin.write(json.dumps(request) + '\n')
    process.stdin.flush()

    # Read response
    response = json.loads(process.stdout.readline())
    return response

# Usage
asyncio.run(call_mcp_tool('snackbase_collections', {'action': 'list'}))

Common Patterns

Natural Language Queries

Transform natural language into SnackBase queries:

User: "Show me active users"
→ snackbase_records tool with filter: { status: 'active' }

User: "Count records by type"
→ snackbase_records with aggregation

Schema-Aware Code Generation

The AI can generate code based on your collection schemas:

User: "Generate a TypeScript interface for the products collection"
→ [Uses snackbase_collections to get schema]
→ Generates TypeScript interface

Data Validation

Use AI to validate data before insertion:

User: "Validate this user data before creating"
→ [Checks collection rules]
→ Validates required fields
→ Checks constraints

Troubleshooting

MCP Server Not Starting

Verify installation: which snackbase-mcp
Check environment variables: echo $SNACKBASE_URL
Test manually: Run snackbase-mcp in a terminal

Tools Not Available

Restart the AI assistant after configuration
Check the MCP server logs for errors
Verify API key has required permissions

Permission Errors

Check API key permissions
Verify account access
Review collection rules if accessing specific data

Best Practices

Security

Use scoped API keys with minimal permissions
Rotate keys regularly
Monitor audit logs for unusual activity
Never expose keys in client-side code

Performance

Use pagination for large result sets
Request only needed fields
Cache frequently accessed data
Use filters to reduce data transfer

Error Handling

try {
  const result = await client.callTool({
    name: 'snackbase_records',
    arguments: { action: 'list', collection: 'posts' },
  });
} catch (error) {
  if (error.message.includes('401')) {
    // API key invalid or expired
  } else if (error.message.includes('403')) {
    // Insufficient permissions
  }
}

Next Steps

Tools Reference - Detailed documentation for each tool

Getting Started

Integration

Tools Reference

​Claude Code Integration

​Setup

​Usage Examples

​Common Workflows

​Data Exploration

​Quick CRUD Operations

​Data Analysis

​Cursor IDE Integration

​Setup

​Usage in Cursor

​Continue.dev Integration

​Setup

​Usage in VS Code

​Cline (Formerly Claude Dev) Integration

​Setup

​Custom Integration

​TypeScript Example

​Python Example

​Common Patterns

​Natural Language Queries

​Schema-Aware Code Generation

​Data Validation

​Troubleshooting

​MCP Server Not Starting

​Tools Not Available

​Permission Errors

​Best Practices

​Security

​Performance

​Error Handling

​Next Steps

Claude Code Integration

Setup

Usage Examples

Common Workflows

Data Exploration

Quick CRUD Operations

Data Analysis

Cursor IDE Integration

Setup

Usage in Cursor

Continue.dev Integration

Setup

Usage in VS Code

Cline (Formerly Claude Dev) Integration

Setup

Custom Integration

TypeScript Example

Python Example

Common Patterns

Natural Language Queries

Schema-Aware Code Generation

Data Validation

Troubleshooting

MCP Server Not Starting

Tools Not Available

Permission Errors

Best Practices

Security

Performance

Error Handling

Next Steps