A Template MCP Server
About
A Template MCP Server is an example implementation of a Model Context Protocol (MCP) server that demonstrates how to connect AI agents and services (like Claude) to external data sources. It uses a PostgreSQL database backend to manage user data and exposes simple tools for user…
Details
- License
- MIT license
Explore
- Defines and exposes custom tools via the MCP standard.
- Handles requests from MCP clients and translates them into database operations.
- Returns structured results or errors to the client.
- Includes four example tools: add, list, find, and delete users.
- Uses async PostgreSQL connection via SQLAlchemy with asyncpg driver.
- Configurable via environment variables (e.g., DB_URL, DB_HOST, DB_PORT).
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
A Template MCP ServerCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Python 3.10+
Access to a PostgreSQL database. The server needs connection details.
Docker (recommended for running the server easily).
uv (for local development dependency management).
python -m venv .venv
source .venv/bin/activate # or .\.venv\Scripts\activate on Windows
uv pip install -e .[dev]
2. Set Environment Variables: Ensure the required DB_URL is set in your shell environment or is present in a .env file in the root directory (dotenv is used).
3. Run the Server:
uv run python sqlmcp/server.py
```
(The server will use host/port from environment variables or defaults)
To illustrate how tools are defined and exposed, this server implements the following example functions for interacting with the user database:
add_user: Adds a new user to the database.
Parameters: name (string), email (string)
Returns: Confirmation message (string) or error.
get_all_users: Retrieves a list of all users from the database.
Parameters: None
Returns: List of user objects (each a dictionary with id, name, email) or error.
find_user_by_email: Finds a specific user by their email address.
Parameters: email (string)
Returns: A single user object (dictionary) or an error if not found.
delete_user_by_email: Deletes a user from the database based on their email address.
Parameters: email (string)
Returns: Confirmation message (string) or error.
These tools serve as concrete examples of the kinds of operations an MCP server might provide access to.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"a template mcp server": {
"agents-mcp-demo": {
"command": "docker",
"args": [
"build",
"-t",
"sql-mcp-demo-server",
"--build-arg",
"PORT=${DB_PORT:-8051}",
"."
]
}
}
}
}McpServers
{
"agents-mcp-demo": {
"command": "docker",
"args": [
"build",
"-t",
"sql-mcp-demo-server",
"--build-arg",
"PORT=${DB_PORT:-8051}",
"."
]
}
}This repository serves as a template, demonstrating how to implement an MCP (Model Context Protocol) server - an open standard designed to connect AI agents and services (like Claude) to tools and data sources. The primary goal of this repository is to illustrate how to build, configure, and run a basic MCP server. The tools are simplified examples and not intended for production use.
This specific example server uses a simple PostgreSQL database backend to manage user data, showcasing how MCP can bridge the gap between AI agents and external resources.
Overview
The server acts as a secure gateway between an MCP client (e.g., an AI agent) and an external data source (in this case, a toy PostgreSQL database containing a users table). It demonstrates how to:
Define and expose custom tools via the MCP standard.
Handle requests from an MCP client.
Translate those requests into operations on the backend resource (e.g., database queries).
Return structured results or errors back to the client.
The demo users table used in this example has the following structure:
id (integer, primary key)
name (text)
email (text, unique)
Example Tools Exposed by the Server
To illustrate how tools are defined and exposed, this server implements the following example functions for interacting with the user database:
add_user: Adds a new user to the database.
Parameters: name (string), email (string)
Returns: Confirmation message (string) or error.
get_all_users: Retrieves a list of all users from the database.
Parameters: None
Returns: List of user objects (each a dictionary with id, name, email) or error.
find_user_by_email: Finds a specific user by their email address.
Parameters: email (string)
Returns: A single user object (dictionary) or an error if not found.
delete_user_by_email: Deletes a user from the database based on their email address.
Parameters: email (string)
Returns: Confirmation message (string) or error.
These tools serve as concrete examples of the kinds of operations an MCP server might provide access to.
Prerequisites
Python 3.10+
Access to a PostgreSQL database. The server needs connection details.
Docker (recommended for running the server easily).
-
uv(for local development dependency management).
Configuration
Environment variables are used to configure the database connection and server settings. Copy the .env.example file to .env and update the values:
```bash
cp .env.example .env
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
