Tag Archives: artificial-intelligence

Introducing Lightweight MySQL MCP Server: Secure AI Database Access

2 Replies


A lightweight, secure, and extensible MCP (Model Context Protocol) server for MySQL designed to bridge the gap between relational databases and large language models (LLMs).

I’m releasing a new open-source project: mysql-mcp-server, a lightweight server that connects MySQL to AI tools via the Model Context Protocol (MCP). It’s designed to make MySQL safely accessible to language models, structured, read-only, and fully auditable.

This project started out of a practical need: as LLMs become part of everyday development workflows, there’s growing interest in using them to explore database schemas, write queries, or inspect real data. But exposing production databases directly to AI tools is a risk, especially without guardrails.

mysql-mcp-server offers a simple, secure solution. It provides a minimal but powerful MCP server that speaks directly to MySQL, while enforcing safety, observability, and structure.

What it does

mysql-mcp-server allows tools that speak MC, such as Claude Desktop, to interact with MySQL in a controlled, read-only environment. It currently supports:

  • Listing databases, tables, and columns
  • Describing table schemas
  • Running parameterized SELECT queries with row limits
  • Introspecting indexes, views, triggers (optional tools)
  • Handling multiple connections through DSNs
  • Optional vector search support if using MyVector
  • Running as either a local MCP-compatible binary or a remote REST API server

By default, it rejects any unsafe operations such as INSERT, UPDATE, or DROP. The goal is to make the server safe enough to be used locally or in shared environments without unintended side effects.

Why this matters

As more developers, analysts, and teams adopt LLMs for querying and documentation, there’s a gap between conversational interfaces and real database systems. Model Context Protocol helps bridge that gap by defining a set of safe, predictable tools that LLMs can use.

mysql-mcp-server brings that model to MySQL in a way that respects production safety while enabling exploration, inspection, and prototyping. It’s helpful in local development, devops workflows, support diagnostics, and even hybrid RAG scenarios when paired with a vector index.

Getting started

You can run it with Docker:

docker run -e MYSQL_DSN='user:pass@tcp(mysql-host:3306)/' \
  -p 7788:7788 ghcr.io/askdba/mysql-mcp-server:latest

Or install via Homebrew:

brew install askdba/tap/mysql-mcp-server
mysql-mcp-server

Once running, you can connect any MCP-compatible client (like Claude Desktop) to the server and begin issuing structured queries.

Use cases

  • Developers inspecting unfamiliar databases during onboarding
  • Data teams writing and validating SQL queries with AI assistance
  • Local RAG applications using MySQL and vector search with MyVector
  • Support and SRE teams need read-only access for troubleshooting

Roadmap and contributions

This is an early release and still evolving. Planned additions include:

  • More granular introspection tools (e.g., constraints, stored procedures)
  • Connection pooling and config profiles
  • Structured logging and tracing
  • More examples for integrating with LLM environments

If you’re working on anything related to MySQL, open-source AI tooling, or database accessibility, I’d be glad to collaborate.

Learn more

If you have feedback, ideas, or want to contribute, the project is open and active. Pull requests, bug reports, and discussions are all welcome.

Scoped Vector Search with the MyVector Plugin for MySQL – Part I

Leave a reply


Semantic Search with SQL Simplicity and Operational Control

Introduction

Vector search is redefining how we work with unstructured and semantic data. Until recently, integrating it into traditional relational databases like MySQL required external services, extra infrastructure, or awkward workarounds. That changes with the MyVector plugin — a native vector indexing and search extension purpose-built for MySQL.

Whether you’re enhancing search for user-generated content, improving recommendation systems, or building AI-driven assistants, MyVector makes it possible to store, index, and search vector embeddings directly inside MySQL — with full support for SQL syntax, indexing, and filtering.

What Is MyVector?

The MyVector plugin adds native support for vector data types and approximate nearest neighbor (ANN) indexes in MySQL. It allows you to:

  • Define VECTOR(n) columns to store dense embeddings (e.g., 384-dim from BERT)
  • Index them using INDEX(column) VECTOR, which builds an HNSW-based structure
  • Run fast semantic queries using distance functions like L2_DISTANCE, COSINE_DISTANCE, and INNER_PRODUCT
  • Use full SQL syntax to filter, join, and paginate vector results alongside traditional columns

By leveraging HNSW, MyVector delivers millisecond-level ANN queries even with millions of rows — all from within MySQL.

Most importantly, it integrates directly into your existing MySQL setup—there is no new stack, no sync jobs, and no third-party dependencies.

Scoped Vector Search: The Real-World Requirement

In most production applications, you rarely want to search across all data. You need to scope vector comparisons to a subset — a single user’s data, a tenant’s records, or a relevant tag.

MyVector makes this easy by combining vector operations with standard SQL filters.

Under the Hood: HNSW and Query Performance

MyVector uses the HNSW algorithm for vector indexing. HNSW constructs a multi-layered proximity graph that enables extremely fast approximate nearest neighbor search with high recall. Key properties:

  • Logarithmic traversal through layers reduces search time
  • Dynamic index support: you can insert/update/delete vectors and reindex as needed
  • Configurable parameters like M and ef_search allow tuning for performance vs. accuracy

Under the Hood: HNSW and Query Performance

MyVector uses the HNSW algorithm for vector indexing. HNSW constructs a multi-layered proximity graph that enables extremely fast approximate nearest neighbor search with high recall. Key properties:

  • Fast ANN queries without external services
  • Scoped filtering before vector comparison
  • Logarithmic traversal through layers reduces search time
  • Dynamic index support: you can insert/update/delete vectors and reindex as needed
  • Configurable parameters like M and ef_search allow tuning for performance vs. accuracy

What’s Next

This post introduces the foundational concept of scoped vector search using MyVector and HNSW. In Part II, we’ll walk through practical schema design patterns, embedding workflows, and hybrid search strategies that combine traditional full-text matching with deep semantic understanding — using nothing but SQL.