Try EnginiQ with Supabase or Postgres in 5 minutes
EnginiQ is the safe Postgres runtime for AI agents. It gives you one guardrailed workflow for schema discovery, migrations, and queries through a Node.js SDK, a CLI, and an MCP server.
Install
Choose the interface that matches your workflow. All three use the same safe Postgres tool surface underneath.
SDK
Embed the engine directly in a Node.js app or internal service.
npm install enginiq-core pg dotenvCLI
Use EnginiQ from your terminal or CI to inspect schema and run migrations.
npm install enginiq-cliMCP
Expose EnginiQ tools to Cursor, Claude, or another MCP-compatible host.
npm install enginiq-mcpConfigure
Point EnginiQ at your Postgres or Supabase database. It resolves configuration in this order: `--database_url`, `ENGINIQ_DATABASE_URL`, `DATABASE_URL`, then `.enginiqrc.json`.
{
"database_url": "postgresql://postgres:postgres@localhost:5432/postgres"
}Run the CLI
npx enginiq doctor
npx enginiq tools
npx enginiq schema
npx enginiq create-table posts
npx enginiq add-column posts title:textTrust controls
EnginiQ now supports preview-first workflows so you can demo or operate more safely before applying changes.
npx enginiq create-table posts --dry-run
npx enginiq add-column posts title:text --read-only
npx enginiq migrate --require-approval
npx enginiq migrate --require-approval --approval-token approved
# Optional audit metadata
npx enginiq create-table posts --dry-run --actor mohan --environment stagingUse the SDK
import {
Engine,
SupabaseConnector,
listTablesTool,
getSchemaTool,
} from "enginiq-core";
const engine = new Engine();
const connector = new SupabaseConnector(process.env.DATABASE_URL);
engine.setConnector(connector);
engine.registerTool(listTablesTool(engine));
engine.registerTool(getSchemaTool(engine));
await engine.connect();
const schema = await engine.runTool("get_schema", {});
console.log(schema);
await engine.disconnect();Available tools
| Tool | Purpose |
|---|---|
| list_tables | Read all public tables |
| describe_table | Inspect one table and its columns |
| get_schema | Read the full public schema |
| create_table | Create a table with guardrails |
| add_column | Safely add a new column |
| run_migration | Apply a named sequence of operations |
| query | Run parameterized SQL with safety checks |
Current guardrails
EnginiQ blocks `DROP DATABASE`, `DROP SCHEMA`, `TRUNCATE`, `ALTER ROLE`, deletes without a `WHERE` clause, and operations on tables prefixed with `auth_`, `storage_`, or `supabase_`.
Mutating workflows can also be run in `dry-run`, `read-only`, or `approval-required` modes for safer previews and review flows.
Frequently asked questions
What is EnginiQ?
EnginiQ is an AI-safe Postgres runtime that gives agents structured tools for schema discovery, migrations, and queries.
Does EnginiQ work with MCP clients?
Yes. The MCP package exposes the same core database tools so IDE agents can call them through a standard protocol.
Which database should I use with EnginiQ today?
Today the product is focused on Postgres-compatible databases, including Supabase-hosted Postgres.