README - Supabase MCP Server by amraly83

Supabase MCP Server

This project implements a Model Context Protocol (MCP) server that provides tools for interacting with a Supabase database. The server enables AI assistants to perform database operations through a standardized interface.

Setup

Clone the repository:

git clone <repository_url>
cd <repository_name>

Install the dependencies:
```
pip install -r requirements.txt
```
The server now accepts Supabase credentials from client requests instead of requiring environment variables. Clients should provide these in their request:

{
    "supabase_url": "your_supabase_url",
    "supabase_key": "your_supabase_service_role_key",
    "tool": "create_table",
    "arguments": {
        "table_name": "users",
        "schema": [
            {"name": "id", "type": "SERIAL PRIMARY KEY"},
            {"name": "name", "type": "TEXT"}
        ]
    }
}

Docker

Build the Docker container:
```
docker build -t mcp-server .
```
Run the container:
```
docker run -p 8000:8000 mcp-server
```
This will build an image named mcp-server and then run it, mapping port 8000 on your host to port 8000 in the container.

GitHub Repository

To push the files to a GitHub repository, follow these steps:

Initialize a Git repository in the current directory:
```
git init
```
Add all the files to the staging area:
```
git add .
```
Commit the changes:
```
git commit -m "Initial commit"
```
Rename the branch to main:
```
git branch -M main
```

Add the remote repository:

git remote add origin https://github.com/amraly83/supabase-mcp-server.git

Push the files to the repository:
```
git push -u origin main
```
You may be prompted for your GitHub username and password. If you have 2FA enabled, you'll need to use a personal access token instead of your password.

Usage

The server provides the following tools:

read_rows(table_name: str = None, limit: int = 10): Reads rows from a Supabase table.
- table_name: The name of the table to read from. This is an optional parameter. If no table name is provided, it will return rows from all tables.
- limit: The maximum number of rows to return. Defaults to 10. This is an optional parameter.
create_record(table_name: str, record: dict): Creates a new record in a Supabase table.
- table_name: The name of the table to create the record in.
- record: A dictionary containing the data for the new record.
update_record(table_name: str, record_id: int, updates: dict): Updates an existing record in a Supabase table.
- table_name: The name of the table to update the record in.
- record_id: The ID of the record to update.
- updates: A dictionary containing the updates to apply to the record.
delete_record(table_name: str, record_id: int): Deletes a record from a Supabase table.
- table_name: The name of the table to delete the record from.
- record_id: The ID of the record to delete.
list_tables(): Lists all tables in the Supabase database.
create_table(table_name: str, schema: list): Creates a new table in the Supabase database.
- table_name: The name of the table to create.
- schema: A list of dictionaries, where each dictionary represents a column in the table. Each dictionary must have the keys "name" and "type".

Example

To use the server, you can send JSON requests to the server's standard input. For example, to read the first 5 rows from a table named "products", you would send the following JSON:

{
    "supabase_url": "your_supabase_url",
    "supabase_key": "your_supabase_service_role_key",
    "tool": "read_rows",
    "arguments": {
        "table_name": "products",
        "limit": 5
    }
}

To list all tables in the Supabase database, you would send the following JSON:

{
    "supabase_url": "your_supabase_url",
    "supabase_key": "your_supabase_service_role_key",
    "tool": "list_tables",
    "arguments": {}
}

To create a new table named "users" with columns "id" (SERIAL PRIMARY KEY) and "name" (TEXT), you would send the following JSON:

{
    "supabase_url": "your_supabase_url",
    "supabase_key": "your_supabase_service_role_key",
    "tool": "create_table",
    "arguments": {
        "table_name": "users",
        "schema": [
            {"name": "id", "type": "SERIAL PRIMARY KEY"},
            {"name": "name", "type": "TEXT"}
        ]
    }
}