Skysql MCP
Launch and delete serverless MariaDB Cloud databases, execute SQL and use database-level AI agents for text-to-SQL and conversations.
What can you do with Skysql MCP?
- Launch and manage serverless MariaDB database instances
- Interact with AI-powered database agents
- Execute SQL queries directly on SkySQL (MySQL/MariaDB) instances
- Manage database credentials and IP allowlists
- List and monitor database services
Install
The installation steps are in the documentation below. Documentation
Documentation
From the project's README on GitHub, refreshed weekly. Links open on GitHub.
SkySQL MCP Server
This package contains everything needed to set up the SkySQL/MariaDB Cloud MCP (Model Context Protocol) server, which provides a powerful interface for managing SkySQL MariaDB database instances and interacting with AI Agents.
Features
- Launch and manage serverless MariaDB database instances
- Interact with AI-powered database agents
- Execute SQL queries directly on SkySQL (MySQL/MariaDB) instances
- Manage database credentials and IP allowlists
- List and monitor database services
Installation
Prerequisites
- Python 3.10 or higher
- A SkySQL/MariaDB Cloud API key
Option 1: Run locally
Installation steps
- Clone the repository:
git clone git@github.com:skysqlinc/skysql-mcp.git cd skysql-mcp - Run the installation script:
chmod +x install.sh ./install.sh - Create a
.envfile in the root directory of the cloned git repository with your SkySQL API key. Obtain API key by signing up for free on SkySQL.SKYSQL_API_KEY=<your_skysql_api_key_here> Start the MCP server (HTTP mode):
chmod +x launch.sh ./launch.shThe server will start on
http://localhost:8000/mcpby default.- Configure your IDE:
Cursor
Add the following to your Cursor MCP config (~/.cursor/mcp.json or .cursor/mcp.json in your project):
{
"mcpServers": {
"skysql-mcp-server": {
"url": "http://localhost:8000/mcp",
"headers": {
"X-API-Key": "<your-skysql-api-key>"
}
}
}
}
The
X-API-Keyheader is sent with each request, allowing per-user API keys when the server is hosted remotely.
Windsurf
Add the following to your Windsurf MCP config (~/.codeium/windsurf/mcp_config.json):
{
"mcpServers": {
"skysql-mcp-server": {
"serverUrl": "http://localhost:8000/mcp"
}
}
}
Note: Windsurf uses
serverUrl(noturl) and cannot send request headers. Start the server in single-tenant mode so it falls back to the API key in your environment:SKYSQL_SINGLE_TENANT=true ./launch.shWith
SKYSQL_API_KEYset in.env(step 3). WithoutSKYSQL_SINGLE_TENANT, the server requires every HTTP caller to supply its ownX-API-Keyheader and will reject requests that don't — deliberately, so a shared deployment can never serve requests using the operator's key.
Claude
Add the server as a custom connector (Settings > Connectors > Add custom connector), then supply your API key under Request headers:
| Field | Value |
|---|---|
| URL | your deployed server URL, ending in /mcp |
| Header name | x-api-key |
| Header value | your SkySQL API key, with no prefix |
| Required | yes |
Generate the key at app.skysql.com/user-profile/api-keys. Enter it exactly as-is — Claude sends the value verbatim and does not add a scheme or prefix.
Claude requires a publicly reachable HTTPS URL, so
http://localhost:8000/mcpwill not work here; deploy the server first. Request header authentication is currently in beta and may need to be enabled for your organization.
- (Optional) Test the server interactively with MCP CLI:
npx @wong2/mcp-cli uv run python src/mcp-server/server.py
Frequently asked questions
What is Skysql MCP?
Launch and delete serverless MariaDB Cloud databases, execute SQL and use database-level AI agents for text-to-SQL and conversations.
How do I install Skysql MCP?
Follow the installation steps in its documentation, shown on this page and in its repository.
Is Skysql MCP open source?
Yes. Its code is on GitHub under the MIT license.
Who maintains Skysql MCP?
It is published by skysqlinc. AI.tn lists it but does not maintain or endorse it.
From the community awesome-mcp-servers lists and the project's GitHub repository. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.