nocentino

nocentino / sql-mcp-server

Public

SQL Server MCP demo - two MCP server patterns on SQL Server 2025: Data API Builder for CRUD and a custom TypeScript DBA monitoring server with 30 DMV-backed tools. Runs in Docker Compose.

www.nocentino.composts2026-05-24-giving-copilot-visibility-into-sql-server-with-mcp data-api-builder dba docker github-copilot mcp
10
2
89% credibility
Found May 27, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

This project is a demonstration environment that lets AI assistants actually 'see' and diagnose SQL Server databases. It sets up two complementary tools: one for exploring your application data (Products, Orders, etc.) through natural conversation, and another with 30 diagnostic tools for database administration tasks like spotting blocking sessions, finding slow queries, checking backup health, and monitoring Always On Availability Groups. The AI never connects to the database directly—it calls these tools instead, which run the queries and return structured results. The AI then synthesizes those results into plain-English explanations. It's designed for database professionals who want to experiment with AI-assisted troubleshooting without writing SQL or risking accidental changes.

How It Works

1
💡 You hear about AI-powered database help

You've been wondering if an AI assistant could actually help you manage your SQL Servers without writing all the queries yourself.

2
📦 You download and launch the demo

You grab the project from the internet and start it up with one command. Everything spins up automatically—your SQL Servers, the AI tools, all of it.

3
🔌 You connect your AI assistant

You point your AI tool (like VS Code Copilot or Claude) to your new server. Now your AI can actually see your databases.

4
🤖 You ask your first question

You type: 'Tell me about this SQL Server—is there anything I should worry about?' The AI immediately starts calling its tools.

5
The AI investigates your problem
⛓️
Blocking investigation

It finds sessions blocking each other, shows you exactly who's blocking whom and for how long

📊
Performance check

It examines wait stats, top queries, and missing index recommendations

💾
Health snapshot

It pulls server info, database status, backup health, and configuration concerns

6
📝 You get a clear diagnosis

The AI explains what it found in plain English—not just raw data, but actionable insights you can understand.

✅ You have answers, not just data

Instead of staring at query results, you have a clear picture of what's happening and what to do next. The AI never touched your database directly—you stayed in control the whole time.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is sql-mcp-server?

This project gives AI assistants like GitHub Copilot and Claude real eyes on your SQL Server estate. It runs two MCP servers in Docker: one built on Microsoft's Data API Builder that exposes application data as REST, GraphQL, and MCP endpoints with zero custom code, and a custom TypeScript server with 30 DBA-focused tools that query SQL Server DMVs directly. The agent never touches the database. It calls the tool server, the tool server runs the SQL, and you stay in control. You can ask "are there any blocking sessions?" and get a diagnosis with the actual blocker SQL, not a query to run manually.

Why is it gaining traction?

The multi-instance support is the hook. One container manages connection pools to multiple SQL Server instances simultaneously, and a single `fan_out_query` tool runs the same T-SQL across your whole fleet in parallel. If one instance is down, the others still return results. The 30 tools cover the full DBA diagnostic surface: blocking chains, wait stats, missing indexes, AG health, backup status, query store regressions, and more. The README walks through real agent scenarios with actual output, showing how Copilot chains multiple tool calls into a single synthesized incident report.

Who should use this?

DBAs who want AI assistants to do actual diagnostic work instead of just writing queries. DevOps teams managing multiple SQL Server instances who need fleet-wide health snapshots. Developers evaluating MCP for database tooling. It's a demo and learning environment, not production infrastructure, but the patterns are production-grade.

Verdict

At 10 stars with minimal community testing, this is an early-stage project with a high credibility score (0.9%) that reflects solid author reputation and documentation quality. The architecture is well-thought-out and the demo scripts make the value proposition tangible. Worth spinning up in a lab to see what agent-driven database monitoring actually looks like in practice.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.