Case study

How Fugle built WatchScanning

Q: Who built WatchScanning?

WatchScanning was designed and developed by Fugle (fugle.no), a Norwegian studio that builds websites, automation systems and AI products for businesses. Fugle delivered the entire product: concept, design, frontend, serverless backend, AI integration, payments and the SEO strategy behind several hundred content pages.

The site you are on right now, an AI service that authenticates and values watches from photos, was designed and developed by Fugle, a Norwegian studio that builds websites, automations and AI products for businesses. This is the story of how the product came together: the choices behind the technology, the architecture and the SEO strategy that drives its traffic.

Updated August 2, 2026 · ~5 min read

What WatchScanning is

The user uploads 2–4 photos of a watch. The AI identifies the brand, model and reference, checks its authenticity against known tells for over 1,000 brands, and estimates its market value. The result is a structured report delivered in seconds, with PDF download, shareable links and follow-up questions in chat. Around the product sits a credit system with Stripe payments, user history, sharing and an admin portal.

From idea to launch in four steps

1

Concept and design

The problem was concrete: the second-hand watch market is flooded with fakes, and a physical authentication costs time and money. The solution had to be fast, visual and trustworthy. The design got a clean, Apple-inspired look with dark and light mode, built to convey precision and trust. When you sell a verdict, trust is the product.

2

The AI core

The analysis uses a multimodal AI model that reads the photos against known tells: typography, engraving, indices, movement details and overall consistency. The hard part was getting the answer back as structured data, with the verdict, reasoning, technical findings and value estimate in a fixed schema, so every report has the same shape and can be validated before the user sees it.

3

The business logic

A scan costs credits, credits are bought through Stripe Checkout, and balances are handled in server-side transactions with automatic refunds if an analysis fails. New users get a free scan, and a referral system rewards both sides. All of it chains together authentication, database and payments without a single traditional server.

4

SEO as the growth engine

Instead of paid ads, Fugle built a content machine: over 600 static guides covering brands, models, serial-number lookups and buying advice, all with structured data and internal linking. The strategy now spans five languages (Norwegian, Danish, Dutch, Polish and Italian) with proper hreflang clusters, winning questions like “er klokken min ekte?” in markets competitors overlook.

The technology under the hood

Layer	Choice	Why
Frontend	React + TypeScript, Vite, Tailwind CSS	Fast development, type safety, minimal framework noise
Backend	Serverless functions on Vercel	Scales automatically, no server maintenance, pay per use
AI	Multimodal image analysis with structured output	Same report shape every time, answers can be validated
Data and auth	Firebase (Google sign-in + Firestore)	Secure login and transaction-safe credit balances
Payments	Stripe Checkout	Proven checkout, card data never touches our systems
Content	600+ static pages, JSON-LD, hreflang in 5 languages	Instant loading and organic traffic without an ad budget

What Fugle builds for businesses

Fugle does not just deliver websites. It builds complete products across four service areas:

Websites

Fast, mobile-friendly websites for businesses and projects, with the SEO foundation in place from day one. This site is the demo.

Automation

Systems that work around the clock and remove manual tasks, from data collection to reporting and alerts.

AI solutions

Custom tools powered by AI: image analysis, document understanding and chat-based services, like the AI core in WatchScanning.

Custom development

Complete products built from scratch: design, frontend, backend, payments and operations under one roof.

Other products from the same workshop: AiStil (AI interior styling with your own furniture), Scrolling.no (micro-learning in a mobile format), Skriving.no (a Norwegian typing test) and BetBott (a football analysis bot with over 66,000 members).

Need something similar for your business?

Whether it is a website, an automation or a complete AI product: describe what you need and you will get a concrete proposal with price and timeline.

Visit fugle.no

Frequently asked questions

Who built WatchScanning?

WatchScanning was designed and developed by Fugle, a Norwegian studio that builds websites, automation systems and AI products for businesses. Fugle delivered the entire product: concept, design, frontend, serverless backend, AI integration, payments and the SEO strategy behind several hundred content pages.

What technology is WatchScanning built with?

The frontend is React with TypeScript, built with Vite and styled with Tailwind CSS. The backend runs as serverless functions on Vercel, with Firebase for authentication and data, Stripe for payments, and a multimodal AI model for the image analysis itself. All content, over 600 static SEO pages in five languages, is served as fast static files.

How much does a website from Fugle cost?

Fugle prices projects individually, because a simple business site and a complete AI product are very different deliveries. Reach out via fugle.no with a description of what you need and you will get a concrete proposal with price and timeline.

Can Fugle build AI solutions for my business?

Yes. WatchScanning is itself an example: a complete AI product with image analysis, structured reports, a credit system and payments. Fugle builds similar solutions for other industries, from chatbots and document analysis to automations that replace manual work around the clock.

What else has Fugle built?

Besides WatchScanning, Fugle has built AiStil (AI interior styling), Scrolling.no (micro-learning in a mobile format), Skriving.no (a Norwegian typing test) and BetBott (a football analysis bot). The common thread is complete products: design, backend, AI and operations under one roof.