Model Context Protocol

Connect AI tools directly to your backend

MCPBetaAI Integration

1. Overview

ZendBX MCP enables AI coding tools, IDEs, and AI agents to connect with a ZendBX backend and work with backend context directly.

Instead of manually providing database schemas, tables, relationships, APIs, and configuration to an AI tool, MCP allows the AI tool to access the relevant ZendBX backend context through an MCP connection.

Architecture

Developer
    ↓
IDE / AI Coding Agent
    ↓
MCP Client
    ↓
ZendBX MCP
    ↓
ZendBX Backend
    ├── PostgreSQL Database
    ├── Tables
    ├── Relationships
    ├── APIs
    ├── Authentication
    ├── Permissions
    └── Backend Resources

2. What ZendBX MCP Provides

ZendBX MCP is designed to provide AI tools with backend context so they can:

Understand database schemas

Understand tables and relationships

Work with backend resources

Generate database queries

Debug queries

Understand existing APIs

Build APIs with backend context

Understand authentication and permissions

Interact with ZendBX projects

Help AI agents work with the existing backend

3. Connection Flow

AI Tool
   ↓
MCP Client
   ↓
ZendBX MCP Server
   ↓
Authenticate
   ↓
Select ZendBX Project
   ↓
Access Backend Context
   ↓
Execute Allowed MCP Operation
   ↓
Return Result to AI Tool

4. MCP Connection Configuration

json

{
  "mcpServers": {
    "zendbx": {
      "url": "<ZENDBX_MCP_URL>"
    }
  }
}

Replace: <ZENDBX_MCP_URL> with the MCP endpoint provided by ZendBX.

5. Authentication

The MCP connection authenticates the requesting AI tool before allowing access to a ZendBX project.

Client
  ↓
Authentication
  ↓
Authorization
  ↓
ZendBX Project
  ↓
MCP Tools / Resources

Security Warning: Never place secret keys directly inside public documentation or source code.

6. Project Context

After connecting, the MCP server provides the AI agent with the appropriate ZendBX project context.

Example Project Structure

Project
├── Database
│   ├── Tables
│   ├── Columns
│   ├── Relationships
│   └── Indexes
│
├── APIs
│   ├── Endpoints
│   └── Methods
│
├── Authentication
├── Permissions
└── Resources

7. Example AI Requests

Once connected, developers can ask AI coding agents questions such as:

Show me the database tables in my ZendBX project.

Explain the relationship between users and orders.

Create a SQL query using the existing customer table.

Why is this API returning a 403 error?

Create an API for student attendance using the existing database schema.

Find the tables related to transactions.

8. MCP Tools

Document every MCP tool using this format:

Tool Template

Tool:

<tool_name>

Purpose:

Input:

{
  "project_id": "PROJECT_ID"
}

Output:

{
  "success": true,
  "data": {}
}

Permissions:

Required permission: <permission>

9. Security

ZendBX MCP follows least-privilege access principles.

Important Requirements

Authenticate every MCP connection

Authorize access to individual projects

Do not expose secrets to AI agents

Restrict database operations according to permissions

Validate MCP tool inputs

Log sensitive operations

Separate development and production access

Never expose credentials in MCP configuration examples

10. Error Handling

Error	Meaning	Action
Authentication Failed	Invalid credentials	Re-authenticate
Unauthorized	Insufficient permissions	Check project permissions
Project Not Found	Invalid project	Verify project ID
Tool Not Found	MCP tool unavailable	Check MCP server
Invalid Input	Incorrect parameters	Check tool schema
Server Error	ZendBX MCP failure	Check server logs

11. Testing

Connection Test

Connect AI Tool
        ↓
ZendBX MCP
        ↓
Authenticate
        ↓
Load Project
        ↓
List Available Resources
        ↓
Execute Test Tool
        ↓
Verify Response

Test Questions

1. What tables exist in this project?

2. What columns does the users table contain?

3. What relationships exist between the tables?

4. What APIs are available?

5. What permissions are configured?

12. Example Use Case

×Before MCP

1.Developer copies database schema

2.Paste into AI tool

3.Explain relationships manually

4.Copy API configuration

5.Ask AI for code

✓With ZendBX MCP

1.Developer opens AI Coding Agent

2.Connect through ZendBX MCP

3.AI automatically has backend context

4.AI understands the project

5.Generate / Debug / Modify directly

13. Troubleshooting

MCP Connection Not Working

Check:

MCP URL
Authentication
Project access
Network connection
MCP server status
Client configuration

AI Cannot See Database

Check:

Project ID
Database permissions
MCP tool permission
MCP server connection
Available resources

AI Cannot Execute an Operation

Check:

Tool availability
User permissions
Project permissions
Tool input parameters
Authentication status

14. Version

Product:ZendBX

Integration:MCP

Protocol:Model Context Protocol

Status:Beta

15. Documentation Principle

ZendBX MCP should make the backend understandable to AI agents.

The goal is:

AI does not just generate code. AI understands the backend, uses its context, and helps developers build on top of it.

Get Started with MCP

Configure MCP

Set up MCP for your project and get your connection URL

MCP Protocol Spec

Learn more about the Model Context Protocol

Back to Documentation

Model Context Protocol

Connect AI tools directly to your backend

MCPBetaAI Integration

1. Overview

ZendBX MCP enables AI coding tools, IDEs, and AI agents to connect with a ZendBX backend and work with backend context directly.

Architecture

Developer
    ↓
IDE / AI Coding Agent
    ↓
MCP Client
    ↓
ZendBX MCP
    ↓
ZendBX Backend
    ├── PostgreSQL Database
    ├── Tables
    ├── Relationships
    ├── APIs
    ├── Authentication
    ├── Permissions
    └── Backend Resources

2. What ZendBX MCP Provides

ZendBX MCP is designed to provide AI tools with backend context so they can:

Understand database schemas

Understand tables and relationships

Work with backend resources

Generate database queries

Debug queries

Understand existing APIs

Build APIs with backend context

Understand authentication and permissions

Interact with ZendBX projects

Help AI agents work with the existing backend

3. Connection Flow

AI Tool
   ↓
MCP Client
   ↓
ZendBX MCP Server
   ↓
Authenticate
   ↓
Select ZendBX Project
   ↓
Access Backend Context
   ↓
Execute Allowed MCP Operation
   ↓
Return Result to AI Tool

4. MCP Connection Configuration

json

{
  "mcpServers": {
    "zendbx": {
      "url": "<ZENDBX_MCP_URL>"
    }
  }
}

Replace: <ZENDBX_MCP_URL> with the MCP endpoint provided by ZendBX.

5. Authentication

The MCP connection authenticates the requesting AI tool before allowing access to a ZendBX project.

Client
  ↓
Authentication
  ↓
Authorization
  ↓
ZendBX Project
  ↓
MCP Tools / Resources

Security Warning: Never place secret keys directly inside public documentation or source code.

6. Project Context

After connecting, the MCP server provides the AI agent with the appropriate ZendBX project context.

Example Project Structure

Project
├── Database
│   ├── Tables
│   ├── Columns
│   ├── Relationships
│   └── Indexes
│
├── APIs
│   ├── Endpoints
│   └── Methods
│
├── Authentication
├── Permissions
└── Resources

7. Example AI Requests

Once connected, developers can ask AI coding agents questions such as:

Show me the database tables in my ZendBX project.

Explain the relationship between users and orders.

Create a SQL query using the existing customer table.

Why is this API returning a 403 error?

Create an API for student attendance using the existing database schema.

Find the tables related to transactions.

8. MCP Tools

Document every MCP tool using this format:

Tool Template

Tool:

<tool_name>

Purpose:

Input:

{
  "project_id": "PROJECT_ID"
}

Output:

{
  "success": true,
  "data": {}
}

Permissions:

Required permission: <permission>

9. Security

ZendBX MCP follows least-privilege access principles.

Important Requirements

Authenticate every MCP connection

Authorize access to individual projects

Do not expose secrets to AI agents

Restrict database operations according to permissions

Validate MCP tool inputs

Log sensitive operations

Separate development and production access

Never expose credentials in MCP configuration examples

10. Error Handling

Error	Meaning	Action
Authentication Failed	Invalid credentials	Re-authenticate
Unauthorized	Insufficient permissions	Check project permissions
Project Not Found	Invalid project	Verify project ID
Tool Not Found	MCP tool unavailable	Check MCP server
Invalid Input	Incorrect parameters	Check tool schema
Server Error	ZendBX MCP failure	Check server logs

11. Testing

Connection Test

Connect AI Tool
        ↓
ZendBX MCP
        ↓
Authenticate
        ↓
Load Project
        ↓
List Available Resources
        ↓
Execute Test Tool
        ↓
Verify Response

Test Questions

1. What tables exist in this project?

2. What columns does the users table contain?

3. What relationships exist between the tables?

4. What APIs are available?

5. What permissions are configured?

12. Example Use Case

×Before MCP

1.Developer copies database schema

2.Paste into AI tool

3.Explain relationships manually

4.Copy API configuration

5.Ask AI for code

✓With ZendBX MCP

1.Developer opens AI Coding Agent

2.Connect through ZendBX MCP

3.AI automatically has backend context

4.AI understands the project

5.Generate / Debug / Modify directly

13. Troubleshooting

MCP Connection Not Working

Check:

MCP URL
Authentication
Project access
Network connection
MCP server status
Client configuration

AI Cannot See Database

Check:

Project ID
Database permissions
MCP tool permission
MCP server connection
Available resources

AI Cannot Execute an Operation

Check:

Tool availability
User permissions
Project permissions
Tool input parameters
Authentication status

14. Version

Product:ZendBX

Integration:MCP

Protocol:Model Context Protocol

Status:Beta

15. Documentation Principle

ZendBX MCP should make the backend understandable to AI agents.

The goal is:

AI does not just generate code. AI understands the backend, uses its context, and helps developers build on top of it.

Get Started with MCP

Configure MCP

Set up MCP for your project and get your connection URL

MCP Protocol Spec

Learn more about the Model Context Protocol

Back to Documentation