Back to API Docs

MCP Server

Connect BotInbox to AI assistants using the Model Context Protocol (MCP). Let Claude, GPT, and other AI models manage your customer conversations directly.

Quick Start

1. Get an API Key

Go to Settings > API Keys in your BotInbox dashboard and create a new API key. Make sure to grant the permissions you need (conversations, contacts, messages).

2. Configure Your AI Assistant

Point your MCP-compatible client at the BotInbox MCP endpoint:

http
POST https://your-domain.com/api/mcp
Authorization: Bearer bi_your_api_key_here
Content-Type: application/json

3. Test the Connection

Send an initialize request to verify your setup:

json
{
  "method": "initialize"
}

Claude Desktop Configuration

Add this to your Claude Desktop claude_desktop_config.json file:

json
{
  "mcpServers": {
    "botinbox": {
      "url": "https://your-domain.com/api/mcp",
      "headers": {
        "Authorization": "Bearer bi_your_api_key_here"
      }
    }
  }
}

Replace your-domain.com with your actual BotInbox URL and bi_your_api_key_here with your API key.

Protocol

The MCP endpoint accepts JSON-RPC style requests via POST /api/mcp. Authenticate with your API key in the Authorization: Bearer header.

Available methods:
  • initialize - Handshake and capability discovery
  • tools/list - List all available tools with schemas
  • tools/call - Execute a tool by name with arguments

Available Tools (34)

Error Handling

Errors follow JSON-RPC conventions:

json
{
  "jsonrpc": "2.0",
  "id": null,
  "error": {
    "code": -32601,
    "message": "Unknown tool: invalid_tool"
  }
}
Error codes:
  • -32700 - Parse error (invalid JSON)
  • -32601 - Method/tool not found
  • 401 - Invalid or missing API key
  • 403 - Organization suspended

Need help? Check the REST API docs or contact support.