The metadata layer for AI-first development.

30-second example

A typed entity in metadata, a project config that picks the generators, and the three CLI commands you actually run. Drift detection is architectural — regenerated code breaks dependent logic at compile time, and meta migrate diffs metadata against the live database.

# metaobjects/subscriber.meta.yaml — the entity definition
metadata.root:
  package: acme
  children:
    - object.entity:
        name: Subscriber
        children:
          - field.long:    { name: id }
          - field.string:  { name: email }
          - field.string:  { name: createdAt, "@autoSet": onCreate }
          - identity.primary: { "@fields": id }

// metaobjects.config.ts — picks generators + DB target
import { defineConfig } from "@metaobjects/cli";
import { entityFile, queriesFile, routesFile, barrel }
  from "@metaobjects/codegen-ts/generators";

export default defineConfig({
  outDir:   "./src/db",
  dialect:  "sqlite",
  generators: [entityFile(), queriesFile(), routesFile(), barrel()],
});

# Scaffold metaobjects/ + .metaobjects/ + metaobjects.config.ts
$ meta init

# Generate code (entities, queries, routes, barrel) per config
$ meta gen

# Diff metadata vs live DB and emit migration SQL files
$ meta migrate --db file:./dev.db --slug add-subscriber

# Flatten loaded metadata to one canonical JSON artifact
$ meta export --out ./snapshot.json

Conformance fixtures (JSON form): fixtures/conformance/. YAML and JSON are equivalent — pick whichever your stack prefers.