lab
Your browser cannot display the WebGPU Synthetic Life canvas.

Press and drag to apply the selected tool

Synthetic Life Foundry

A living-image instrument for growing behavior, not animating a preset shape.

Generation0000
Genomebioluminal reef
Active toolFeed
Frame rate—

Waiting for browser capability check.

The image is not animated. It evolves.

Most animation follows a plan: an object is told where to move and when to arrive. Synthetic Life Foundry begins with simple rules for food, growth, competition and decay, then lets each new image grow from the one before it.

You choose the conditions and apply a tool, but you do not draw every branch. The luminous forms are a record of many small exchanges accumulating over time. That makes the experience feel cultivated rather than choreographed.

  1. 01

    Seed

    Begin with the same spread of food and active material each time.

  2. 02

    Feed

    Add resources that make new growth possible near the pointer.

  3. 03

    Compete

    Nearby colonies consume resources and change each other’s surroundings.

  4. 04

    Mutate

    Introduce variation so the field can break symmetry and find new shapes.

  5. 05

    Divide

    Growth turns into new edges, chambers and branching paths.

  6. 06

    Decay

    Older regions fade, release space and make room for the next form.

Each new picture begins with the picture before it.

Images in memory
02
Kinds of information
04
Nearby points checked
09
Local target
60 fps

The browser keeps two copies of the field. It reads the current one, writes the changed result into the other, then swaps them. This simple back-and-forth gives the image memory without asking the main page to recalculate every point one by one.

A second drawing step turns those stored values into light and color. Small differences between neighboring points create tissue-like edges, while each genotype supplies its own color family. The same field can be viewed as the whole colony, its membrane or its remaining nutrient.

For curious readers: the two copies are floating-point GPU textures, and their exchange is often called a ping-pong feedback loop. The 60 fps figure is a local Chromium observation, not a promise for every device.

The result is fed back into the next moment.

The system repeats these five steps many times each second.

  1. 01RememberPrevious image

    Begin with the field left by the last update.

  2. 02Look nearbyLocal neighborhood

    Check the current point and the eight points around it.

  3. 03Grow or competeColony + food

    Exchange material and decide whether this area grows or recedes.

  4. 04Accept your inputSelected tool

    Apply feeding, mutation, erasing, flow or renewal near the pointer.

  5. 05Draw the next imageUpdated field

    Save the result and turn its edges and density into visible light.

You shape the conditions instead of drawing every outcome.

The presets are not recorded videos and the controls are not filters. Each one changes a behavior: how fast the field evolves, how much growth costs, how strongly colonies compete, how long recent activity remains or how the environment moves.

This creates a useful balance between control and surprise. A genotype gives the work a recognizable character, while feedback keeps the exact performance from feeling pre-recorded.

This is a visual ecology, not a living organism.

  • It borrows ideas from spreading, competition and resource exchange; it is not a scientific model of life.
  • The field has memory and can form surprising patterns, but it has no awareness or intention.
  • The full experience requires WebGPU. Unsupported browsers receive an authored static fallback.
  • The prototype is locally verified Lab work; it is not proof of a released public product.

Each point stores four simple values.

Those four values travel and change together, giving the whole image a compact memory.

  1. RFood base

    Material that can be exchanged as the field changes.

  2. GColony

    How much active growth is present at this point.

  3. BNutrient

    A renewable resource that growth gradually consumes.

  4. ARecent growth

    A fading trace of where division has just happened.

One system can grow in three distinct ways.

  1. 01Bioluminal ReefBalanced growth

    Rounded chambers, luminous edges and a steady exchange of resources.

  2. 02Neural MyceliumSearching network

    Fast spread and variation produce thin, branching paths.

  3. 03Predator BloomCompetitive growth

    High appetite carves abrupt fronts through the surrounding field.

A growing image could become a flexible media system.

These are research ideas, not finished features. New signals could influence the ecology while the same recognizable visual character remains intact.

  1. Proposed studyGenerative motion

    Create recognizable visual families that still grow differently in every performance.

  2. Proposed studyAudioreactive ecology

    Let rhythm, bass and musical changes influence fertility, appetite, variation and flow.

  3. Proposed studyData-shaped growth

    Let changing values affect the environment instead of drawing them as a conventional chart.

  4. Proposed studyResponsive identity

    Give a campaign a recognizable visual character without repeating one fixed animation.