Database schema

Launch Now stores everything in PostgreSQL through Drizzle ORM. The whole schema lives in a single file, drizzle/auth-schema.ts: the Better Auth tables, billing plans, feedback, notifications and the example resources.

Where the schema lives

import { drizzle } from "drizzle-orm/node-postgres"
import { Pool } from "pg"
import * as schema from "./auth-schema"
 
export const db = drizzle(
  new Pool({
    connectionString: process.env.DATABASE_URL,
    connectionTimeoutMillis: 30_000,
  }),
  { schema }
)

Better Auth reads and writes the same tables through its Drizzle adapter (drizzleAdapter(db, { provider: "pg", schema }) in lib/auth.ts).

To change the schema, edit drizzle/auth-schema.ts, then generate and apply a migration:

pnpm db:generate
pnpm db:migrate

See Scripts for every database command.

Conventions

• Primary keys are text IDs. Better Auth generates them for its tables; app code uses randomUUID().
• Column names are snake_case in SQL and camelCase in TypeScript (createdAt maps to created_at).
• timestamp columns are timestamp without time zone.
• updatedAt columns use $onUpdate(() => new Date()), so Drizzle sets them on every update.
• Foreign keys to user and organization use ON DELETE CASCADE unless noted.

Authentication tables

These tables are required by Better Auth and its plugins (admin, organization, API key, two-factor, Stripe, database rate limiting). Keep their column names in sync with Better Auth if you upgrade it.

user

session

Index: session_userId_idx on user_id.

account

One row per sign-in method linked to a user (OAuth provider or credential for a password).

Indexes: unique account_issuer_accountId_uidx on (issuer, account_id), and account_userId_idx on user_id.

verification

Short-lived values such as email OTP codes.

Index: verification_identifier_idx on identifier.

rate_limit

Storage for Better Auth's rate limiter (rateLimit.storage: "database" in lib/auth.ts).

apikey

API keys created by the API key plugin.

Indexes: apikey_configId_idx, apikey_referenceId_idx and apikey_key_idx.

two_factor

Indexes: twoFactor_secret_idx on secret and twoFactor_userId_idx on user_id.

session_log

History of ended sessions, shown in the admin user detail page next to active sessions (getUserById in lib/admin-actions.ts) and used to detect sign-ins from a new device in lib/auth.ts.

Index: sessionLog_userId_idx on user_id.

Rows are written by the database, not by app code: migration drizzle/0003_petite_grim_reaper.sql creates a log_session_deleted() function and a session_delete_trigger trigger that copies each session row into session_log before it is deleted.

Organization tables

Created for Better Auth's organization plugin.

organization

member

Indexes: member_organizationId_idx and member_userId_idx.

invitation

Indexes: invitation_organizationId_idx and invitation_email_idx.

Billing tables

subscription

Written by the Better Auth Stripe plugin. reference_id is either a user ID or an organization ID, depending on who owns the subscription.

plan

The plan catalog, seeded by pnpm db:seed:plans (free, pro, ultra).

Index: plan_sortOrder_idx on sort_order.

Feedback

feedback

Messages sent from the in-app feedback dialog (createFeedback in lib/feedback-actions.ts).

Indexes: feedback_userId_idx and feedback_status_idx.

Notification tables

See Notifications for how these tables are used.

notification

One row per notification. The audience column decides who sees it:

• user: the user in user_id.
• organization: members of organization_id, optionally only those whose role is in roles.
• admins: platform admins (user.role = 'admin').
• all: every user (broadcasts).

Reads and archives are stored in notification_receipt, so a broadcast stays a single row.

Indexes: notification_userId_idx on (user_id, created_at), notification_organizationId_idx on (organization_id, created_at) and notification_audience_idx on (audience, created_at).

notification_receipt

Per-user read and archive state.

Primary key: (notification_id, user_id). Index: notification_receipt_userId_idx.

notification_preference

Per-user, per-category delivery settings. A missing row means the category defaults from NOTIFICATION_CATEGORIES in lib/notifications/events.ts apply.

Primary key: (user_id, category).

Example resource tables

These tables back the example CRUD pages. They sit between @reset-remove:start and @reset-remove:end markers, so pnpm reset-project removes them from the schema. Run pnpm db:generate and pnpm db:migrate afterwards to drop them.

categories

User-scoped example.

Index: categories_userId_idx.

projects

User-scoped example.

Indexes: projects_userId_idx and projects_categoryId_idx.

org_projects

Organization-scoped example, generated from drizzle/resources/org-projects.json by pnpm resource:generate.

Indexes: org_projects_userId_idx and org_projects_organizationId_idx.

Relations

Drizzle relations (declared with relations() in drizzle/auth-schema.ts) power the with option of db.query. They do not change the SQL schema.

The subscription, plan, apikey, verification, rate_limit and notification tables have no Drizzle relations. Query them with explicit joins or filters.

const users = await db.query.user.findMany({
  columns: safeUserColumns,
  with: {
    sessions: {
      columns: { id: true, createdAt: true, updatedAt: true, expiresAt: true },
      orderBy: desc(session.updatedAt),
      limit: 1,
    },
  },
  orderBy: desc(user.createdAt),
})

Resource specs

drizzle/resources/ does not contain tables. It holds JSON specs for the resource-kit CRUD generator, plus a README.md that documents every option. The generator turns a spec into a table appended to drizzle/auth-schema.ts, server actions, pages and a manifest entry in lib/resource-kit/manifest.json.

{
  "name": "orgProject",
  "title": "Projects",
  "scope": "organization",
  "icon": "FolderKanban",
  "labelField": "name",
  "fields": [
    { "name": "name", "type": "text", "required": true, "searchable": true },
    { "name": "description", "type": "textarea" },
    {
      "name": "status",
      "type": "select",
      "options": ["planned", "active", "paused", "done"],
      "defaultValue": "planned"
    },
    { "name": "dueDate", "type": "date" }
  ]
}

This spec produces the org_projects table above: the camelCase name becomes a plural snake_case table name, each field becomes a column, and the generator adds id, user_id, created_at and updated_at. The organization scope adds organization_id with a cascading foreign key and an index.

To add your own table from a spec:

pnpm resource:generate --spec drizzle/resources/your-resource.json
pnpm db:generate
pnpm db:migrate

See pnpm resource:generate for the flags and field types.