Mcp Mysql
Let your coding agent inspect production or dev MySQL/MariaDB schema and run read-only SQL without opening a GUI or risking writes.
Overview
mcp-mysql is a MCP server for the Build phase that gives AI agents read-only, SQL-guarded access to MySQL and MariaDB for queries, schema discovery, and health checks.
What is this MCP server?
- Read-only SQL execution with SQL-level guarding against mutating statements
- List and describe tables for schema exploration from Claude Code or Cursor
- MySQL and MariaDB connectivity via stdio MCP (npm @infoinlet/mcp-mysql v0.1.1)
- Health tooling so agents can sanity-check connectivity before deep queries
- Marketplace subfolder services/mcp-mysql from infoinlet-marketplace
- Server version 0.1.1 on npm as @infoinlet/mcp-mysql
- Transport: stdio only per server.schema.json
- Repository subfolder: services/mcp-mysql in infoinlet-marketplace
What problem does it solve?
You need the agent to reason over real table structures and query results, but you cannot safely hand it a full SQL client with write permissions.
Who is it for?
Indie devs with an existing MySQL/MariaDB instance who want schema-aware agent answers during backend work or read-only production checks.
Skip if: Teams that need migrations, writes, or admin DDL from the agent—this server is explicitly read-only.
What do I get? / Deliverables
After you register the stdio server and credentials, the agent can describe schemas and run guarded read-only SQL from the chat instead of context-switching to a DB client.
- Agent-callable tools for table list/describe, guarded queries, and DB health
- Read-only SQL path without exposing write statements to the model
- Repeatable stdio MCP wiring from the infoinlet marketplace package
Recommended MCP Servers
Journey fit
Backend and data modeling work is where agents most often need live table metadata and safe SELECTs while you ship features. The server targets relational backends—list/describe tables, health checks, and guarded queries map directly to backend integration and debugging.
How it compares
MCP database connector with read-only guards, not a migration skill or ORM code generator.
Common Questions / FAQ
Who is mcp-mysql for?
Solo builders and small teams using Claude Code, Cursor, or similar agents who already run MySQL or MariaDB and want safe, read-only database context in the agent loop.
When should I use mcp-mysql?
Use it when debugging queries, exploring schema, or verifying data during backend build or operate triage, and you want SQL-guarded SELECTs instead of write access.
How do I add mcp-mysql to my agent?
Install @infoinlet/mcp-mysql from npm (v0.1.1), add a stdio MCP entry in your agent config pointing at the package, and pass read-only database host, user, and password via environment variables your setup documents.