SeeRepo with Next.js

SeeRepo recognizes a Next.js app from the next package in its package.json. It works with the App Router and the Pages Router, and with monorepos: each package is read with its own framework.

What SeeRepo reads

App Router

• Every page file is a screen.
• Every route file is an action. Routes under api/cron/ are things that run on their own.
• Webhooks are API doors for machines: SeeRepo spots them by their path, or by the signature check inside them (Stripe's constructEvent, Svix, Shopify's HMAC). So are .well-known, an api/mcp route, the sitemap, robots and Open Graph images.
• A file that starts with 'use server' is an action: your server actions.
• middleware and instrumentation are doors too, and layout, loading and error files are special files of the router.

Pages Router

• Files under pages/ are screens, and pages/api/ are actions.
• _app and _document are doors.

Who comes in

A screen that requires a session is for Signed-in people; one that doesn't is for Visitors. Webhooks come from Machines.

An example from our test bank

vercel/nextjs-subscription-payments, a Next.js App Router starter with Supabase and Stripe, read at a fixed commit:

• 4 screens, 4 actions and 1 API door (app/api/webhooks/route.ts, Stripe's webhook)
• 5 data models from its Supabase migration: customers, prices, products, subscriptions and users
• Services found with their proof: Supabase (8 files) and Stripe (7 files)
• 3 screens for visitors and 1 that requires a session (app/account/page.tsx)

Good to know

• Data models are read from Prisma, Supabase migrations and Mongoose. Drizzle models aren't read yet: in our bank, nextjs/saas-starter uses Drizzle and its map shows no data models.
• Services come from three signals: the packages you import, the names of your environment variables and the URLs in your code.
• See also Supabase, Prisma and Mongoose.