Jordanics Hanna and Galántai Gábor built the Napsarló project in the days before the August 12 partial solar eclipse in Budapest: an online shop for eclipse glasses created and automated with artificial intelligence (Claude Opus 5). Over roughly four days they sold nearly 30,000 glasses, generated gross revenue above 100 million HUF, and expect 40–45 million HUF in pre‑tax profit after costs.
How it started: shortage and a quick decision
The initiative began when the pair discovered that eclipse glasses were unavailable in Budapest ahead of the August 12 eclipse. Galántai Gábor — cofounder of CoinCash — used AI to locate a nearby shop with 25 glasses in stock; they bought them and sold them within hours. On August 7 they were in Austria, which accelerated their immediate buying decision.
They contacted the wholesaler and were told that 200 units could be collected immediately, which allowed them to scale the operation.
The webshop and the role of AI
They built the webshop using the Claude Opus 5 model: the site napfogyatkozasszemuveg.shop was operational within hours and provided order management, inventory tracking, card payments and options for home delivery or in‑person pickup. The AI also produced first drafts of the general terms and conditions and the privacy policy; these drafts were reviewed and approved by a lawyer.
Card payments were finally implemented via Stripe, connected to their company Superioritas Kft., after registrations with Simple and Barion proved too slow for the tight timeline.
Sales figures and financials
They priced the glasses at 3,990 HUF each and offered discounted 2‑ and 4‑pack deals. Nearly 30,000 units were sold; gross revenue exceeded 100 million HUF, which corresponds to roughly 80 million HUF net revenue before VAT. After expenses they expect a pre‑tax profit of 40–45 million HUF. They spent about 8 million HUF on Google ads.
At peak demand they received 732 orders in a single hour for more than 1,500 glasses; by the end of the project almost 30,000 pairs had been sold.
Logistics: the biggest challenge
Orders arrived faster than two people could process, so they recruited temporary staff. Using AI they developed an admin interface to onboard workers and track tasks (enveloping, labeling, sticker application) and inventory. Initially they printed labels on a home printer, then moved to a print shop and bought sticker paper for faster processing.
Home deliveries were handled by GLS with bulk shipments; personal pickups created queues of thousands of orders. Using AI analysis, they selected three Budapest pickup points (Óbudai-sziget, Erzsébet tér, Móricz Zsigmond körtér) and built a booking system with QR‑code verification. This method enabled them to hand over more than 12,000 pairs in a single day.
Operations and administration
The project’s tax affairs were handled through Superioritas Kft., and invoicing was tied to Billingó; documentation provided to Forbes.hu verifies the number of transactions. The webshop automated invoicing and customer communications (email, SMS).
What next?
Initially the Napsarló project was conceived as a one‑off; the shop closed ordering and now shows a memorial page where buyers can upload photos and messages from the eclipse. The day after the interview, Galántai Gábor called to say they are considering developing Napsarló further after the unexpected success, but offered no details.
Why this matters
The Napsarló case demonstrates how AI can rapidly assemble a full e‑commerce operation — website, legal documents draft, payment integration, logistics and HR modules — in response to a short‑term demand spike. It also highlights operational and logistical challenges when scaling quickly with a small team, and how targeted advertising and automation can enable rapid sales growth.
Source: interview statements by Jordanics Hanna and Galántai Gábor and documentation provided to Forbes.hu.
