Configuration

Manual configuration

If you prefer to configure manually, create jsorm.config.ts at the root:

import { defineConfig } from "@jsorm/core";

export default defineConfig({
  models: "src/models",
  databases: {
    main: {
      access: "server",
      provider: "postgres-node",
      options: {
        naming: { fields: "snake_case" },
      },
    },
  },
});

Options

defineConfig

Option	Type	Description
models	string	Relative path from the project root where models are located
databases	Record<string, DatabaseConfig>	Available databases

DatabaseConfig

Option	Type	Description
access	"server" | "client"	Model available on server or client
provider	string	Provider name (e.g., pg-node, sqlite-node, indexeddb)
options	Record<string, any>	Provider-specific options
options.naming.fields	"snake_case" | "camelCase"	Naming convention for columns

Naming

naming.fields	JS property	DB column
(not set)	emailVerified	emailVerified
snake_case	emailVerified	email_verified
camelCase	emailVerified	emailVerified

If naming is not configured, DB columns will use the exact same name as the JS property (no transformation applied). For example, a property named emailVerified maps directly to the column emailVerified.

Table names are explicit — defined directly in defineModel().

Environment variables

Use env() from @jsorm/core to mark values as env vars:

import { defineConfig, env } from "@jsorm/core";

export default defineConfig({
  databases: {
    main: {
      provider: "sqlite-node",
      options: { database: env("DATABASE_URL_SQLITE") },
    },
  },
});

.env

DATABASE_URL_SQLITE=./data/prod.db

If the value is not an env var name (e.g., ./data.db or :memory:), it is used directly without resolution.