WildSide

Step into the WildSide. Explore living worlds inspired by real places, meet their small inhabitants, and discover the nature hiding in everyday surroundings.

Inspiration

A backyard looks small until you imagine how many lives unfold inside it. A tree is someone’s shelter. A patch of grass is an entire landscape. The animals we glimpse for a second carry on living after we look away.

We wanted to make those lives feel closer. Nature is part of familiar streets, courtyards and small patches of water as much as distant wilderness. Caring can begin with noticing a creature, recognizing its home and wanting to return. WildSide invites people to step inside those places and see neighbors where they once saw scenery.

What it does

WildSide is a multiscale exploration prototype: enter a living place, change your perspective and discover the inhabitants that make it feel alive. Walk at human scale, explore from a bird’s perspective or move through an insect-sized landscape. Follow a bird, notice a beetle, encounter a spider and its web, or watch a water strider at the surface.

The current prototype brings together a courtyard, two habitat interpretations, a real map-based Kyoto block with 36 buildings and 15 street plans, and a bounded, wet Everglades underwater experience. These places offer different ways to explore scale, movement and life.

Save meaningful viewpoints locally and return to them. Revisit current or historical geometry, browse memories across places, and refine or remove the sources behind a world. Optional synthesized above-water sound and rain add atmosphere. Observations and interpreted simulation remain distinct, so an imagined encounter does not become a claim that an animal was recorded there.

Automatic capture and depth reconstruction are implemented in development work now being integrated into the standard experience. The reviewed photo and video entry path currently combines calibration and interpreted landmarks; video can suggest distinct views and track a previously reviewed ground patch.

How we built it

The current app uses Three.js and JavaScript modules, with a Node.js 22 local/offline workflow. Public geography and custom habitat geometry supply places to explore; custom behaviors give their small inhabitants movement and presence.

A shared, source-aware place model connects geometry, navigation, inhabitants, observations and persistent memories. This lets a corrected source improve a place while keeping saved experiences connected to it. Local transfer and recovery support continuity between visits.

We are also experimenting with an opt-in HAMON WebGPU rendering path. It is separate from the standard Three.js experience and remains experimental.

Challenges we ran into

The hardest problem was coherence. A world must keep its scale, surfaces, movement and inhabitants aligned as the visitor changes perspective. A bird’s route, an insect’s contact with the ground and a person’s saved viewpoint all need to belong to the same place.

Real places are also incomplete and change over time. We needed source corrections and removal to coexist with memories and historical geometry, without presenting interpreted details as observations. Integrating strong individual components into the experience people actually open proved just as important as building those components.

Visual quality demanded watching the world in motion. Automated checks can catch regressions, but convincing contact, readable movement and a sense of life still require looking closely at the experience.

Accomplishments that we're proud of

We joined multiscale navigation, real map-based geography, small inhabitants and source-linked memories in one prototype. The result gives visitors reasons to look more closely: a bird to follow, a web to notice, a water surface to investigate and a place worth returning to.

We are especially proud of making continuity part of exploration. A saved moment can remain meaningful as the world’s geometry and sources are refined, instead of every visit starting from an unrelated scene.

What we learned

A place starts to feel alive through relationships: between a creature and its surface, a visitor and their viewpoint, or a memory and the source that anchors it. More geometry alone does not create that connection.

We also learned to judge features in the complete experience. Correct data and passing checks matter, but scale changes, movement and visual clarity have to work together for curiosity to turn into attachment.

What's next for WildSide

We are bringing the existing automatic capture work into the standard experience and strengthening semantic placement, so personal photographs and videos enrich a continuing world with less manual setup. We also want stronger original art, richer motion and rendering, broader habitats and deeper continuity as places change.

The ambition is for someone to close WildSide, look outside, and see neighbors where they previously saw scenery.