GitHub Repository

Chat with your database. Locally. VS Code extension for exploring and understanding schemas and stored procedures using local LLMs. SQL Server · PostgreSQL · MySQL. Zero cloud dependency.

5 starsTypeScript

SchemaSight – Chat with your database schema locally using Ollama

Name: SchemaSight – Chat with your database schema locally using Ollama
Availability: InStock
Author: HiteshShinde26

by HiteshShinde26·Mar 8, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidNiche GemSolve My ProblemBig Brain

Local-first schema Q&A for VS Code, but ChatGPT, Cursor, and Continue already handle this via cloud LLMs.

Strengths

•Zero cloud dependency with local Transformers.js embeddings (all-MiniLM-L6-v2) + Ollama means schema data never leaves the machine—genuine security/compliance win for regulated industries.
•RAG pipeline over schema metadata + object definitions is well-scoped; extension architecture (guided onboarding, model-switching panel, context visibility) feels intentional.
•Solves real legacy-database pain: hours lost reading stored procedures; specific to SQL Server (rare in open source).

Weaknesses

•RAG + local LLM for schema Q&A is already covered by Cursor (with Claude/GPT backend) and Continue (MCP protocol); local-only is a feature, not a moat.
•GitHub README shows ~2 stars, no visible user feedback, no npm/VS Code Marketplace install count; unclear if extension is actually published or still in development.

Post Description

Built this to solve a problem at my job — legacy databases full of stored procedures where understanding existing business logic takes days of reading SQL.

SchemaSight connects to SQL Server, PostgreSQL, or MySQL, crawls your schema, summarizes every object with a local LLM, and lets you ask questions in plain English from inside VS Code. Everything runs on your machine — no cloud, no API keys, no data leaving your environment.

Happy to answer questions about the RAG pipeline or architecture.