QueryMind
Overview
Secure AI-powered Text-to-SQL engine with real-time execution.
QueryMind is an end-to-end Text-to-SQL platform that translates natural language into validated PostgreSQL queries using Gemini 2.5 and FastAPI backend. It features DML/DDL injection prevention and real-time SSE-based streaming dashboard.
Context
Business users struggle to write SQL queries safely. Existing Text-to-SQL systems are prone to injection risks and unreliable outputs.
Obstacles Faced
Preventing DML/DDL injection risks
Ensuring SQL validation before execution
Streaming LLM responses efficiently
Handling large database context
Engine Solutions
- Built security layer eliminating injection risks
- Used Fernet encryption for credential safety
- Implemented SSE streaming in Next.js dashboard
- Used Gemini 2.5 for SQL generation
Outcomes & Impact
Enabled safe automated query generation
Reduced manual SQL dependency
Secure multi-database connection management
System Architecture
Multi-database connection management
Vector storage via Pinecone
Dockerized deployment
FastAPI backend with SSE streaming
Execution Lessons
"Secure AI system design"
"Database query validation pipelines"
"Vector storage and retrieval"
"SSE streaming implementation"