Release v0.1.4 (What’s new?).
Welcome to mcp_ohmy_sql Documentation¶
👀 Overview¶
mcp_ohmy_sql is a powerful SQL Model Context Protocol (MCP) server that bridges AI assistants with your databases. Built on SQLAlchemy’s robust foundation, it provides universal database connectivity with intelligent query optimization, configurable access controls, and built-in safeguards against excessive data loads to LLMs.
Transform your database interactions with natural language queries, automatic schema discovery, and intelligent result formatting—all while maintaining enterprise-grade security and performance.
🚀 Feature Highlights¶
Universal Database Support
Connect to virtually any SQL database through SQLAlchemy’s proven architecture. From lightweight SQLite to enterprise PostgreSQL, MySQL, Oracle, and SQL Server—all supported out of the box.
- Multi-Database Architecture
Manage multiple databases and schemas simultaneously from a single MCP server. Perfect for complex environments with dev/staging/production databases or multi-tenant applications.
- Intelligent Query Optimization
Built-in query analysis engine prevents expensive operations, automatically limits result sets, and provides performance feedback to help you write efficient queries.
- AI-Optimized Schema Encoding
Schema information is compressed by ~70% using a specialized encoding format, dramatically reducing token usage while preserving all essential metadata for accurate query generation.
- Enterprise-Ready Security
Fine-grained table filtering, parameterized query support, and read-only operations by default. Access controls ensure your production data stays safe.
💎 Why Choose mcp_ohmy_sql?¶
While other SQL MCP servers exist, mcp_ohmy_sql stands out through:
- ✨ Comprehensive Database Ecosystem
Beyond traditional SQL databases, we’re expanding to support modern data platforms including AWS Aurora, Redshift, Glue Catalog, MongoDB Atlas SQL, ElasticSearch, OpenSearch, DuckDB, and S3 data files.
- 🔧 Production-Ready Architecture
Designed for real-world usage with connection pooling, error handling, query timeouts, and result size limits that prevent your LLM conversations from breaking.
- 📊 Intelligent Result Formatting
Query results are automatically formatted as Markdown tables—the optimal format for LLM comprehension, using 24% fewer tokens than JSON while maintaining perfect readability.
- 🔒 Security-First Approach
Built-in safeguards include SQL injection prevention, read-only operations, table filtering, and upcoming fine-grained access controls for enterprise deployments.
- 🎯 Developer Experience
Comprehensive documentation, clear error messages, and extensive configuration options make setup and maintenance straightforward.
Coming Soon: Remote MCP server deployment, advanced access controls, and expanded database ecosystem support.
See our ROADMAP.md for the complete vision and upcoming features.
🚀️ Supported Features¶
See our ROADMAP.md for the complete vision and upcoming features.
Feature |
Status |
Note |
|---|---|---|
Multi Database Support |
✅ Supported |
|
Local MCP Server via UV |
✅ Supported |
|
Local MCP Server via Docker |
⏳ In Progress |
|
Remote MCP Server |
⏳ In Progress |
|
One Click to Deploy Remote MCP Server |
⏳ In Progress |
|
Export Results to Local Files |
⏳ In Progress |
|
Local Data File Analysis |
⏳ In Progress |
|
User Management |
⏳ In Progress |
Remote MCP server only feature |
Access Control Management |
⏳ In Progress |
Remote MCP server only feature |
🛢️ Supported Databases¶
See our ROADMAP.md for the complete vision and upcoming features.
Database |
Status |
Note |
|---|---|---|
Sqlite |
✅ Supported |
via Sqlalchemy |
Postgres |
✅ Supported |
via Sqlalchemy |
MySQL |
✅ Supported |
via Sqlalchemy |
Oracle |
✅ Supported |
via Sqlalchemy |
MSSQL |
✅ Supported |
via Sqlalchemy |
AWS Aurora |
⏳ In Progress |
via boto3 |
AWS Redshift |
✅ Supported |
via boto3 |
AWS Glue Catalog Databases |
⏳ In Progress |
via boto3 |
MongoDB |
⏳ In Progress |
via Atlas SQL |
ElasticSearch |
⏳ In Progress |
via ElasticSearch SQL |
OpenSearch |
⏳ In Progress |
via OpenSearch SQL |
DuckDB |
⏳ In Progress |
via duckdb |
Data Files on AWS S3 |
⏳ In Progress |
via boto3 |
🎯 Get Started¶
Quick Start Guide: Set up and run the server in under 5 minutes