Configuration

Baeta is configured through a single file at the project root.

Configuration file

Baeta looks for a file named baeta or .baeta with any of these extensions, in the working directory:

• .ts, .mts
• .js, .mjs

note

The config file must be ESM. For .ts/.js, set "type": "module" in package.json; otherwise use the .mts or .mjs extension.

Minimum config

1. Point at your schemas

import { defineConfig } from "@baeta/cli";

export default defineConfig({
  graphql: {
    schemas: ["src/**/*.gql"],
  },
});

2. Wire up package.json scripts

{
  "scripts": {
    "build": "baeta generate",
    "start": "baeta generate --watch --run='node --watch --inspect src/app.ts'",
    "types": "tsc --noEmit"
  }
}

What these do:

• build — generates GraphQL types and module typedefs from your schema files.
• start — runs the generator in watch mode and launches the app alongside it via --run. The --run process is long-lived: Baeta starts it once and only respawns it if it exits. Live reload comes from node --watch inside --run, which restarts the app when source files (including the typedefs Baeta regenerates) change.
• types — typechecks without emitting. Use it in CI.

tip

If you're using Bun or Deno you can replace start with either:

{
  "start:bun": "baeta generate --watch --run='bun --watch --inspect src/app.ts'",
  "start:deno": "baeta generate --watch --run='deno --watch --inspect --allow-net --allow-read --allow-env src/app.ts'"
}

tip

Run baeta generate --help for the full flag list. --watch re-runs the generator on schema changes. --run='…' is launched once as a long-running process; Baeta won't kill it on each regenerate, so the inner command (typically node --watch / bun --watch / deno --watch) handles its own reloads.

Next

Create your first module. For the full set of config options, see the Configuration Reference.