VOXA is a fictional bank of the future. We built its site on 28 September 2026 as a concept for banks and fintech products, in the genre of interactive feature stories where a preloader counts to ten and a camera flies through one 3D world, a feature per stop — and rebuilt everything inside that format: the models, the screens, the films and the features themselves.

The brief we set ourselves: a bank you talk to and spend time in. A voice replaces the card number, friends replace the list of payees, and a trip, a dream or a split bill becomes something people do together — the bank as your second place online after the messenger.

The world is one three.js scene with ten dioramas in a pastel studio whose colour flows from stop to stop; the camera flies between them along an arc, with a slight lens breathing, and follows the cursor. Every model is built in code: phones with live screens drawn on canvas, a white plane circling a boarding pass, a glass jar filling with coins, metal, glass and holographic cards, rising hearts, an arch carrying the members of an investment circle. The menu shows thumbnails rendered from the same scene when the page loads.

The voice works for real in the browser. Speech is turned into text by the browser’s own recognition service and parsed into an intent in English or Russian — transfer, request, message, dream deposit, flight, card, split, coach question — including Russian word forms such as «Ане» and «Максу». Voice ID records three phrases and builds a voiceprint from the average spectral shape and pitch, computed in the tab; later commands are checked against it before the money “moves”, and a mismatch asks for Face ID instead. The whole site can be steered by voice too: “next”, “back”, “menu”, “show the dream”.

Each feature opens a working phone app on one shared state, so money sent in one window appears in the chat, the balance and the test lab. Friends reply and react, the dream jar fills while you watch, a bought ticket becomes a boarding pass with a live countdown. The closing scene cuts two short films generated with Veo 3.1 — Alex asks VOXA to pay Anna for pizza, Anna laughs when it lands — together with the two phones on the page, then hands the microphone to the visitor. Everything is in English and Russian.

The concept is one page with ten features, ten phone demos, a test lab and the closing scene, in English and Russian. A parser run on 28 September 2026 understood all 56 test commands, 28 in each language. On the same day we counted frames with requestAnimationFrame for three seconds, twice resting on a slide and twice during a camera flight, in headless Chrome with ANGLE Metal on a MacBook with an Apple M5 at 1600 × 1000 on a Retina scale: 60.1, 59.8, 60.0 and 58.1 fps.

The page loads 0.24 MB of JavaScript with gzip, three.js included. The whole build is 4.3 MB, 3.3 MB of which are the four film clips (English and Russian versions of each).