What is the benefit of Result Normalization?

Result Normalization provides consistent JSON formatting for all query results, complete with metadata, paging, and truncation markers, making data consumption by LLMs reliable and predictable.

How does SQL ensure safe query execution for LLMs?

It implements robust safety features including READ-only by default, customizable table allow/deny lists, query timeouts, and result row limits to prevent unintended or malicious operations.

Which database engines does SQL currently support?

SQL offers a unified access layer for PostgreSQL, MySQL/MariaDB, and SQLite, with planned support for other major databases like Snowflake, BigQuery, and SQL Server.

SQL

Name: SQL
Author: waabox

bywaabox

Gestión de Bases de Datos

Desarrollo de API

Herramientas para Desarrolladores

Provides a unified, secure layer for LLMs and agent workflows to execute and manage SQL queries across various databases.

Acerca de

This experimental tool provides a minimal and fast server designed to expose SQL databases to Large Language Models (LLMs) and agent workflows via a clean, typed, and controllable protocol. It aims to unify SQL access across multiple database engines, making it LLM-friendly, safe, auditable, and easily pluggable into various automation pipelines. By offering features like schema introspection, safe query execution, and result normalization, it ensures predictable and secure interaction with critical SQL data sources.

Características Principales

Unified SQL access layer for multiple database engines (PostgreSQL, MySQL, SQLite)
Schema Introspection (tables, columns, types, foreign keys, views, sample rows)
Safe Query Execution (READ-only by default, allow/deny lists, timeouts, row limits)
Query Explanation Mode (EXPLAIN, ANALYZE, plan summaries)
Result Normalization (consistent JSON formatting with metadata, paging, truncation)
0 GitHub stars

Casos de Uso

Providing a unified interface for database operations across multiple SQL engines
Enabling LLMs to safely query and interact with SQL databases
Integrating secure SQL access into agent workflows and automation pipelines

Acerca de

Características Principales

Unified SQL access layer for multiple database engines (PostgreSQL, MySQL, SQLite)
Schema Introspection (tables, columns, types, foreign keys, views, sample rows)
Safe Query Execution (READ-only by default, allow/deny lists, timeouts, row limits)
Query Explanation Mode (EXPLAIN, ANALYZE, plan summaries)
Result Normalization (consistent JSON formatting with metadata, paging, truncation)
0 GitHub stars

Casos de Uso

Providing a unified interface for database operations across multiple SQL engines
Enabling LLMs to safely query and interact with SQL databases
Integrating secure SQL access into agent workflows and automation pipelines