The Power Of AI In Creating Immersive Depth Experiences At Abyssal Station

The FABLE/175 exhibition has made Abyssal Station, its sixth room, available as a single-page web experience that converts scrolling into a simulated 3,800-meter ocean descent. The exhibition says the site was built end-to-end by AI, making it a test of how generated code and art direction can produce a coherent, interactive environment rather than a conventional page.

Abyssal Station links the visitor’s scroll position to a master depth value. That value controls the depth meter, background color, lighting and pressure display, while also influencing particle movement and animated sea life. The result is designed to make every part of the interface respond to the same simulated location below the surface.

The experience moves through several ocean zones. Its design brief calls for schooling fish in the sunlight zone, physics-driven jellyfish in the twilight layer, an anglerfish and marine snow in the midnight zone, and ghostly amphipods near the trench. At the bottom, station lights activate as a restrained finale.

According to the published build account, the site uses HTML, CSS and JavaScript without frameworks, content delivery networks or external image requests. Visuals are generated through CSS, SVG and canvas code. The creators also report reduced-motion behavior, keyboard navigation, visible focus states and self-hosted fonts, although the source does not include independent test results for those features.

One Depth Model Unifies the Experience

The central design choice is not any single animation but the use of one simulated depth coordinate to govern many systems at once. That approach ties together color, light, interface data and creature behavior, reducing the risk that separate visual effects will feel disconnected.

For designers and developers, Abyssal Station offers a concrete example of AI-assisted work aimed at interaction design and atmosphere, not only text or static imagery. It also shows how a detailed art-direction prompt can specify performance limits, responsive layouts and accessibility requirements alongside visual style. The finished room may help readers judge where AI-generated front-end work is effective and where human testing remains necessary.

Inside the FABLE/175 Web Exhibition

Abyssal Station is Room 6 of 175 in FABLE/175, which is described by its publisher as a completed exhibition of fundamentally different AI-built websites. Other rooms include a solar observatory, a living herbarium and an animated type foundry, with each room built around a separate visual and technical premise.

The exhibition credits the original Abyssal Station brief to Claude Fable 5 in an art-director role and says it was executed through the FABLE/175 pipeline. The documented process had three stages: an initial build and self-critique, an external critique intended to identify at least 10 problems, and a final art-direction pass. The source does not identify the external reviewer or provide the full revision history.

“The page is a descent.”

— The published Abyssal Station art-direction brief

Performance and Authorship Lack Proof

The available material does not provide measured frame rates, accessibility audit results or browser coverage. A target of 60 frames per second and layouts for 390-pixel, 834-pixel and 1,440-pixel widths appear in the brief, but requirements are not evidence of achieved performance.

The source also does not document how much human editing or debugging occurred after the AI generated the site. Its description of an end-to-end AI build comes from the exhibition itself and has not been independently verified. The exact model configuration, prompt history, development time and code revision record are also absent, limiting comparisons with human-led or mixed production methods.

Live Testing Is the Next Check

The immediate next step is testing the live room across the screen widths and input methods named in its brief. Independent checks could establish whether the site meets its stated frame-rate, contrast, keyboard and reduced-motion goals.

FABLE/175 is also continuing its room-by-room documentation of the completed exhibition. Future reports may show whether the three-pass production method produces similar results across other concepts and whether the publisher releases more detailed build evidence.

Key Questions

What is Abyssal Station?

Abyssal Station is Room 6 of the FABLE/175 web exhibition. It is a single-page interactive site that represents scrolling as a descent through ocean zones to a fictional deep-sea station.

How does the simulated descent work?

A master scroll anchor converts the visitor’s position into a simulated depth. The site uses that value to update lighting, color, pressure data, particles and creature animation.

Was the entire website made by AI?

The exhibition says the room was built end-to-end by AI from a detailed art-direction brief. The source does not disclose the full extent of human review, code correction or production oversight, so that claim cannot be independently measured from the supplied material.

Does the experience use external images or frameworks?

According to its published brief, the room uses no frameworks, CDNs or external image assets. Its visuals are produced with HTML, CSS, SVG, JavaScript and canvas rendering.

Is Abyssal Station accessible on mobile and with reduced motion?

The creators report support for responsive layouts, keyboard navigation and reduced-motion preferences. No independent audit was included, leaving the site’s real-world accessibility and device performance unconfirmed.

Source: Thorsten Meyer AI