The app was developed end-to-end in about three hours using Goose.
- Initial scaffolding created with
create-next-app
. - All other changes were vibe coded using Goose—from
git init
tovercel deploy --prod
. I didn’t edit a single line of code directly, and I haven’t read most of the lines in this project. - OpenAI models used:
o1
,o3
, ando4-mini
. - Total development cost (API calls): < $50. I experimented with different models; cost could be < $5 if using
o4-mini
exclusively.
I didn’t save all prompts used during development, but most commits represent the work done from a single prompt (with occasional follow-ups). Future commits will include prompts in the commit messages.
This Next.js app displays who plans to be in the office each day, using Google SSO (via NextAuth.js) and fetching data from the Forkable API.
- Google SSO restricted to a specific email domain (
cleanlab.ai
). - Server- and client-side route protection.
- Forkable API integration with a 30-day cached session cookie.
- Responsive dashboard UI with modern styling.
Install dependencies and run the development server:
npm install
npm run dev
Open http://localhost:3000 in your browser.
Create a .env.local
file in the project root with the following:
GOOGLE_CLIENT_ID=...
GOOGLE_CLIENT_SECRET=...
NEXTAUTH_SECRET=...
ALLOWED_EMAIL_DOMAIN=cleanlab.ai
FORKABLE_ADMIN_EMAIL=...
FORKABLE_ADMIN_PASSWORD=...
FORKABLE_CLUB_IDS=123,456
Deploy on Vercel or any Node.js hosting platform. Ensure your environment variables are set in your deployment settings.