MCP server integration

Full control over AI integration. The Model Context Protocol (MCP) allows you to expose your own tools, prompts, and resources directly to the AI assistant. Your MCP server defines what actions the AI can perform and what data it can access - giving you complete flexibility over the integration.

Full ControlCustom ToolsYour Server
Tool Definition
{
"name": "create_booking",
"description": "Create a new taxi booking",
"inputSchema": {
"type": "object",
"properties": { ... }
}
}
Tool Response
{
"content": [{
"type": "text",
"text": "Booking #12345 created"
}]
}

Tools

MCP

Tools are the primary way your MCP server extends the AI assistant's capabilities. Each tool represents an action the AI can perform - from querying your database to creating bookings or sending notifications. The AI decides when and how to use these tools based on the conversation context.

Custom actions

Define any action you need - check vehicle availability, validate addresses, calculate fares, manage bookings. Your tools, your business logic, your rules.

Typed parameters

Each tool declares its input schema. The AI understands what parameters are required, their types, and how to gather them from the conversation naturally.

Rich responses

Return structured data that the AI can interpret and present to users. Include success/error states, messages, and any business data needed.

Example: Book Taxi Tool

Tool Input Schema
{
"type": "object",
"properties": {
"pickup": { "type": "string" },
"destination": { "type": "string" },
"time": { "type": "string" }
},
"required": ["pickup"]
}
Tool Call Result
{
"content": [{
"type": "text",
"text": "Booking created.\nID: 12345\nETA: 8 min"
}]
}

Prompts

MCP

Prompts are reusable templates that guide the AI's behavior in specific scenarios. Define how the AI should handle common situations - greeting returning customers, explaining pricing, or handling complaints. Prompts ensure consistent, high-quality responses aligned with your brand.

Conversation templates

Create templates for common scenarios. The AI uses these as guides while maintaining natural conversation flow and adapting to specific situations.

Dynamic arguments

Prompts can accept arguments to customize their behavior. Pass customer names, booking details, or any context needed for the specific interaction.

Brand voice

Define your company's communication style. Whether formal or friendly, technical or simple - ensure the AI represents your brand consistently.

Example: Greeting Prompt

Prompt Definition
{
"name": "greet_customer",
"description": "Greet returning customer",
"arguments": [{
"name": "customerName",
"required": true
}]
}
Prompt Message
{
"messages": [{
"role": "user",
"content": {
"type": "text",
"text": "Welcome back, John!"
}
}]
}

Resources

MCP

Resources provide the AI with access to your data. Expose customer profiles, booking history, pricing information, or any other data the AI needs to assist customers effectively. Resources are read-only and can be updated in real-time as your data changes.

Customer data

Give the AI access to customer profiles, preferences, and history. Enable personalized service without exposing your entire database.

Business information

Share pricing tables, service areas, operating hours, and policies. The AI can answer questions accurately using your actual data.

Real-time updates

Resources reflect your current data. When vehicle availability changes or prices update, the AI immediately has access to the latest information.

Example: Customer Resource

Resource List
{
"resources": [{
"uri": "customer://1234",
"name": "Customer Profile",
"mimeType": "application/json"
}]
}
Resource Content
{
"contents": [{
"uri": "customer://1234",
"text": "{\"name\": \"John\", ...}"
}]
}

Getting started

Setup

Setting up your MCP server is straightforward. Implement the MCP protocol, define your tools, prompts, and resources, then provide us with your server endpoint. We handle the connection and make your capabilities available to the AI assistant.

Protocol implementation

MCP uses JSON-RPC 2.0 over various transports. SDKs are available for TypeScript, Python, and other languages to simplify implementation.

Server configuration

Provide your MCP server endpoint in the dashboard. We establish a secure connection and discover your available tools, prompts, and resources.

Testing & monitoring

Test your integration in our sandbox environment. Monitor tool calls and responses in real-time through the dashboard.

Join our community to get weekly updates and unique gifts every Friday.
Taxi Copilot
© 2026 Taxi Copilot. All rights reserved.