Show HN: Guardrail Layer – self-hosted AI data layer for secure DB chat

2 weeks ago 2

Your database’s new best friend 🧠🔐


Guardrail Layer is an early-stage, open‑source, self‑hosted backend that sits between your database and any tool (LLM, automation, or dashboard).
It enforces data privacy, access rules, and redaction automatically — so you can safely query and expose data without leaking sensitive fields.
Think of it as your data’s very own bodyguard, keeping things safe while you focus on building cool stuff.

(Yep, it’s early but working — and we’d love your feedback to make it even better!)


Dashboard Screenshot Connection Validation UI Query Redaction Example

(More demos and GIFs coming soon!)


  • 🔒 Automatic Redaction — mask emails, IDs, or sensitive values at query time
  • 🧠 Schema Awareness — AI‑ready metadata for smarter, safer queries
  • 💬 Natural‑Language Queries — ask “Who are our top customers by revenue?”
  • ⚙️ Connection Validation — test Postgres/MySQL access through a clean UI
  • 🐳 Self‑Hosted — run anywhere via Docker Compose
  • 🧱 Extensible — built with Node.js, Prisma, TypeScript, and Postgres

Layer Technology
Frontend React, TypeScript, TailwindCSS, Vite
Backend Node.js, Express, Prisma
Database PostgreSQL
Deployment Docker Compose

git clone https://github.com/tyoung1996/guardrail-layer.git cd guardrail-layer

(Grab some ☕ — you’re about to make your data way safer!)

Then open http://localhost:5173 in your browser.
(Watch the magic happen — early but already pretty slick!)

3️⃣ Connect Your Database

Go to the Connections tab and enter credentials for your PostgreSQL or MySQL instance.
The system validates, introspects your schema, and applies safe default policies automatically.
(Your data deserves a bodyguard — and that’s exactly what you’re setting up!)


Goal Guardrail Layer Enables
Connect internal DB to ChatGPT / Local LLM Redact sensitive columns & query safely
Build an analytics chatbot Translate natural language → SQL with guardrails
Expose a read‑only API Enforce row‑level access & policy‑based filters

We’re early, we’re eager, and we want your help!
This project is still evolving, and your feedback can shape the future.

  1. Fork the repo
  2. Create a branch: git checkout -b feature/my-feature
  3. Commit changes and open a Pull Request

(Community vibes incoming — Discord & GitHub Discussions launching soon!)
Jump in, say hi, and help us build something awesome together.
https://discord.gg/tDuPDAeypR


Licensed under the Apache License 2.0.

© 2025 Tyler Young


Star the repo to follow new features and updates — let’s build this fun journey together!

Read Entire Article