Schemabrain

Name: Schemabrain
Author: Arun-kc

@Arun-kc

About Schemabrain

A read-only trust + intelligence layer between AI agents and your database — the agent never writes SQL, PII is refused before the query runs, and every call lands in a tamper-evident audit log. Postgres today.

#mcp #database #security #postgres #pii #data-governance #audit #semantic-layer

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "schemabrain": {
      "command": "uvx",
      "args": [
        "schemabrain",
        "serve",
        "--url-env",
        "DATABASE_URL"
      ],
      "env": {
        "DATABASE_URL": "postgresql://user:password@localhost:5432/dbname"
      }
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Schemabrain?

Schemabrain is an MCP server that acts as a read-only trust and intelligence layer between AI agents (Claude, Cursor, Windsurf) and your SQL database. It compiles SQL from definitions you control and refuses queries touching blocked categories before the database is queried.

How to use Schemabrain?

Install Schemabrain using the uvx package manager. It sits between an AI agent and a Postgres database. Twelve read-only MCP tools are available; the agent never writes SQL directly.

Key features of Schemabrain

Read-only architecture — 12 MCP tools, none can write.
PII-aware refusal — blocked-category queries are rejected before hitting the database.
Tamper-evident audit — SHA-256‑chained append‑only log of every call, refusal, and recovery.
Postgres support today; Apache‑2.0 licensed.
Simple installation via uvx.