This codebase is a demonstration of the principles defined in my Becoming a 1000x Developer blog series.

What makes this codebase special?

When you try to use AI to iterate on this codebase, you're going to find that the AI is just so good at implementing features. This is because of the structure.

For the best results, you should:

• Use Cursor or Claude Code (it will load the cursor rules which are very important to get the most out of this project). You can use another agentic editor, just be sure to transfer the rules (I will add rule files for Windsurf and Claude Code 🔜).
• Use Claude 4 Opus. It is by far, the best AI code model. OpenAI o3 is also good, but it is very slow (imho).

This project is a bun monorepo with the following core components:

• packages/api: The API with services defined using the repository pattern.
• apps/http-server: The HTTP server with tRPC endpoints. Deployable to Cloudflare Workers.
• apps/expo-app: The Expo app.
• apps/next-app: The Next.js application.
• packages/db: The database schema and migrations.
• packages/db-migrations: Migrations for the database, with tooling to apply them using Drizzle.
• packages/models: Data models for the app.
• packages/tooling: The shared tooling for the project.
• utils/seed-data: Tooling to seed the database.

To run the project, you need to have:

• docker (download here) installed and running
• task (instructions here)
• bun (download here)

Once you have those dependencies installed, you can initialize the .env files with:

task initialize-env-files

This will copy the .env.example files to the correct locations (.env, utils/seed-data/.env, packages/db/.env, apps/next-app/.env, apps/http-server/.env).

You can start the dependencies with:

Note (convenience): you can also use this command to destroy and start the docker containers, and apply migrations (all in one)

task docker:reset

Or, you can start the dependencies with:

task docker:start # start the docker containers
task db:migrate # apply the migrations

Then, you can start the project (api, db, and next-app) with:

task start

You can access all services at the following URLs:

• Server: http://localhost:3042
• tRPC: http://localhost:3042/trpc
• Drizzle Studio: https://local.drizzle.studio/?port=8082
• SwaggerUI: http://localhost:8081
• OpenAPI HTTP server: http://localhost:3043
• Next.js app: http://localhost:3000

You might be wondering "why is there a SwaggerUI and an OpenAPI server?"

This is for convenience, and to demonstrate two different transport layers. You can call all of the services through plain-old HTTP client, or through tRPC (which is how the Next.js app calls the API).

This flexibility demonstrates the power of the tRPC library and provides a useful GUI (through SwaggerUI) for the API.

You can login to the app by visiting http://localhost:3000/login and using the following credentials:

• Admin: admin@starterp.com / password
• User: user@starterp.com / password

You can seed the database with:

task seed-db

When you run task docker:reset, the seeding is automatically run.

The utils/seed-data/index.ts file contains the script which performs the seeding.

The Expo app is a mobile app built with React Native and Expo.

You can run the app with:

cd apps/expo-app
bun run start

This will start the app in the Expo Go app.

You can lint all the code with:

task lint:fix:type-check

You can follow the instructions in the DEPLOYMENT.md file.