ajan-sql
An MCP server that provides safe, read-only SQL access for AI agents to query databases (PostgreSQL, MySQL, SQLite) with schema awareness and guardrails.
README
<p align="center"> <img src="./assets/cover.svg" alt="ajan-sql cover" width="860" /> </p>
<p align="center"> <b>Safe, read-only SQL access for AI agents β via MCP.</b> </p>
<p align="center"> <a href="https://www.npmjs.com/package/ajan-sql"><img src="https://img.shields.io/npm/v/ajan-sql" /></a> <a href="https://www.npmjs.com/package/ajan-sql"><img src="https://img.shields.io/npm/dt/ajan-sql" /></a> <a href="https://borakilicoglu.github.io/ajan-sql/"><img src="https://img.shields.io/badge/docs-vitepress-5d98ea" /></a> <a href="./LICENSE"><img src="https://img.shields.io/badge/license-MIT-lightgrey" /></a> </p>
β‘ What is ajan-sql?
ajan-sql is an MCP server that lets AI agents safely query your database.
π read-only
π schema-aware
π guardrailed
Supports:
- PostgreSQL
- MySQL
- SQLite
π Quick Start
DATABASE_URL=postgres://USER:PASSWORD@HOST:PORT/DB npx ajan-sql
Thatβs it.
π§ What it solves
AI agents querying databases is risky.
Without guardrails:
- they can modify data
- run heavy queries
- break your system
π ajan-sql fixes this by enforcing strict rules.
π₯ Safety by default
All queries are:
SELECTonly- no
INSERT,UPDATE,DELETE - no
DROP,ALTER,TRUNCATE - limited results (
LIMIT 100) - timeout enforced (max 5s)
- no multi-statement queries
- no SQL comments
π These rules cannot be bypassed.
π Sandboxing & Approvals
ajan-sql enforces sandboxing at the SQL access layer:
- every query is validated before execution
- query execution is readonly and bounded
- results are limited by row count, timeout, and size
- optional schema/table access policies can restrict readable tables
Human approval flows are handled by the MCP host or client. ajan-sql does not provide its own approval UI.
For production use, connect with a database user that only has readonly permissions. Database permissions should be the final safety backstop.
β‘ Available Tools
list_tablesdescribe_tablelist_relationshipssearch_schemarun_readonly_queryexplain_querysample_rowsserver_info
π§ Example
{
"tool": "run_readonly_query",
"arguments": {
"sql": "SELECT * FROM users LIMIT 10"
}
}
Tool responses include both standard MCP structuredContent and an embedded text/toon version of the same payload.
π Supported Databases
# PostgreSQL
DATABASE_DIALECT=postgres
DATABASE_URL=postgres://USER:PASSWORD@HOST:PORT/DB
# MySQL
DATABASE_DIALECT=mysql
DATABASE_URL=mysql://USER:PASSWORD@HOST:PORT/DB
# SQLite
DATABASE_DIALECT=sqlite
DATABASE_URL=file:/absolute/path/to/database.sqlite
Optional readonly policy controls:
AJAN_SQL_ALLOWED_SCHEMAS=public,analytics
AJAN_SQL_ALLOWED_TABLES=public.users,analytics.events
AJAN_SQL_DENIED_TABLES=public.audit_logs
AJAN_SQL_AUDIT_LOG=true
βοΈ Features
- MCP-native SQL access
- multi-database support
- strict read-only guardrails
- schema discovery + introspection
- structured JSON output for AI agents
- TOON-formatted embedded tool results
- predictable execution limits
- type-safe schemas
π§ Use Cases
- AI copilots querying databases
- internal data assistants
- analytics agents
- safe DB access in automation
- MCP-based workflows
π¦ Install
npm install -g ajan-sql
or:
npx ajan-sql
π§© Client Example
{
"mcpServers": {
"ajan-sql": {
"command": "npx",
"args": ["ajan-sql"],
"env": {
"DATABASE_DIALECT": "postgres",
"DATABASE_URL": "postgres://USER:PASSWORD@HOST:PORT/DB"
}
}
}
}
π‘ Philosophy
AI should never have unsafe database access.
ajan-sql ensures queries are safe, predictable, and controlled.
β€οΈ Support
If this tool helps you:
β Star the repo
β Support via GitHub Sponsors
https://github.com/sponsors/borakilicoglu
π Links
- GitHub: https://github.com/borakilicoglu/ajan-sql
- npm: https://www.npmjs.com/package/ajan-sql
- Docs: https://borakilicoglu.github.io/ajan-sql/
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.