Database Migrations & Types

Plan AI utilizes a perfectly synchronized, end-to-end type safety pipeline.

Any time you make a change to the database schema, that change must be propagated up through the backend controllers, out to the Swagger API specification, and finally injected into the front-end clients (Web, Desktop, and Mobile).

The Golden Rule: yarn update

If you modify the database schema located at plan-ai/backend/prisma/schema.prisma, you must not run standard Prisma migration commands manually.

Instead, always run the master update script from the root of the repository:

yarn update

What does yarn update do?

When you run yarn update, the following automated pipeline triggers:

1. Database Migration: Runs prisma migrate dev to update your PostgreSQL schema.
2. Prisma Client Generation: Generates the new Node.js Prisma Client.
3. Controller Compilation: Uses tsoa to parse your Express controllers and strictly typed request/response interfaces.
4. Swagger Generation: Compiles a new swagger.json OpenAPI specification file.
5. Frontend Syncing: Triggers openapi-typescript inside the Web App, Native Desktop App, and Mobile App. This automatically writes a new api.d.ts file in all three frontends.

Because of this pipeline, the moment you add a new column to the database, it instantly becomes available with full autocomplete in your React components.

Deploying Migrations to Production

When you are deploying a new version of Plan AI to a production server (where you do not want to wipe the database), do not run yarn update.

Instead, use the standard deployment command:

cd plan-ai/backend
yarn prisma:deploy

This safely applies the existing SQL migration files to your production database without generating new migrations.