🛰️ Steampipe MCP Server
@sunil-kumar-h
This MCP server translates the user query to a steam pipe query that could be run on downstream cloud to fetch information on resources
Overview
What is Steampipe MCP Server?
A prototype MCP (Model Context Protocol) server that integrates with Steampipe and AWS, enabling users to query AWS cloud resources using natural language. It translates plain‑English questions into Steampipe SQL queries, making cloud insights accessible to developers, analysts, and non‑technical users.
How to use Steampipe MCP Server?
Clone the repository (git clone https://github.com/sunil-kumar-h/steampipe-mcp-server.git), configure your Steampipe and AWS credentials, then run the server to connect it with MCP‑compatible AI tools like Claude.
Key features of Steampipe MCP Server
- Natural language interface for plain‑English queries
- Dynamic SQL generation from user input
- Query EC2, S3, IAM, CloudTrail, and more
- Retrieve AWS cost and usage data
- Evaluate compliance and security configurations
- Built on FastMCP for MCP protocol support
Use cases of Steampipe MCP Server
- Count running EC2 instances with a simple question
- Identify publicly accessible S3 buckets
- Retrieve highest AWS cost from last month
- Check for IAM users with administrator access
- Scan environments for security and compliance gaps
FAQ from Steampipe MCP Server
What kind of questions can I ask?
You can ask about AWS resources in natural language, such as “How many EC2 instances are running?” or “Which S3 buckets are publicly accessible?”
Do I need to know SQL or AWS terminology?
No. The server translates plain English into Steampipe SQL queries, so you don’t need to understand AWS‑specific terms or write raw SQL.
What AWS services are supported?
The server can query EC2, S3, IAM, CloudTrail, and AWS billing tables, among others supported by Steampipe.
Is this ready for production?
No, this is a prototype. It is intended for experimentation and development use.
How do I set up the server?
Clone the repository, ensure Steampipe and your AWS credentials are configured, then run the MCP server (no additional install commands are provided beyond cloning).