Both Next.js products are full products, not demos. The framework carries real data, more than one language and an AI agent at work.
Hiki’s front end is Next.js 15 with React 19 and the App Router. It is one of two services on Google Cloud: the Next.js app renders the catalog, and a FastAPI back end scrapes the bookstores, enriches every title and serves the data. Scrapers run every day, and change detection flags new arrivals, price drops and stock changes. Both services run on Cloud Run, and Cloud Build deploys them on every push to main.
The catalog is multilingual from the routing up. next-intl handles locale-aware routes, genre names, tag translations and the interface, so readers browse in their own language, titles included. They filter by 21 genres, 100+ tags, price range, discount, stock and source store, and full-text search works in every supported language.
The interface uses a Neubrutalism design system: bold 2px borders, offset shadows and a manga-inspired look, built on shadcn/ui components themed to match.
Scout is also Next.js 15 with the App Router and React 19, but here the app steers an AI agent. The agent discovers companies, enriches contacts, scores each lead from 1 to 5 against the ideal customer profile and drafts personalized outreach. Top-scoring leads then seed a second wave of discovery.
Next.js hosts everything the user touches: a seven-step setup wizard from the ICP to the first lead batch, a drag-and-drop Kanban pipeline, an analytics dashboard and an activity log of every agent action. While the agent runs its cycles, the interface updates live, so progress shows as it happens. The data sits in SQLite in WAL mode, a zero-config, file-based database.
In both products Next.js is one layer of something we designed, built and shipped end to end, from design to deployment, with no handoffs.