A living AI world where biology, computation, platforming, learning, and buildcraft become one persistent artificial life.

ALife is a 3D radial-gravity platforming game and on-device artificial-life system built around three characters:

• Proto — primordial biological life that discovers animal forms and cleansing powers.
• Kee — robotic/computational life that discovers gadgets and real computer-science algorithms.
• ALife / You — the persistent synthesis of the two, able to build lasting combinations from what has been discovered.

The game takes inspiration from the compact, playful traversal of Super Mario Galaxy-style planetary platforming and the long-term buildcraft/endgame loop of action RPGs such as Diablo, but the progression model is built around capabilities rather than character levels.

The central loop is:

Explore → discover → traverse → cleanse → unlock → build → master → automate → push further

Core World Structure

ALife is divided into multiverses.

Each multiverse contains seven planets.

The important distinction is that the planets are not simply seven conventional levels.

The planets are destinations. The space between them is the platformer.

A campaign progresses roughly like this:

Planet 1
   │
   ├── Planet-bridging platforming route
   ▼
Planet 2
   │
   ├── Planet-bridging platforming route
   ▼
Planet 3
   │
   ├── Planet-bridging platforming route
   ▼
Planet 4
   │
   ├── Planet-bridging platforming route
   ▼
Planet 5
   │
   ├── Planet-bridging platforming route
   ▼
Planet 6
   │
   ├── Final planet-bridging challenge
   ▼
Planet 7 / Boss

The player can see and travel through a physical planetary system rather than selecting isolated stages from a menu.

Planet-Bridging Levels

The actual authored platforming levels are the journeys between planets.

These routes use ALife's radial-gravity platforming system and can contain:

• curved strips of terrain,
• tiny planets,
• large planets,
• partial planetary surfaces,
• walls that become floors,
• changing curvature,
• gravity transitions,
• floating structures,
• nested planetary spaces,
• launch sequences,
• traversal gadgets,
• animal abilities,
• infected machinery,
• and other dynamically constructed geometry.

A "planet" does not need a complete spherical walkable surface. A level can use only the curved land required for traversal while still producing the sensation of moving around a planetary body.

This allows the campaign to push radial platforming far beyond a single conventional spherical world.

Target length

A normal planet-bridging challenge should generally be compact:

Approximately three minutes for a successful run.

The goal is dense mechanical variety rather than long stretches of filler.

What Happens on a Planet?

Planets act as persistent campaign nodes.

They can provide:

• discovery,
• environmental storytelling,
• animals,
• Proto powers,
• Kee technology,
• algorithms,
• infected robots,
• resources,
• campfires,
• secrets,
• build preparation,
• and the setup for the next journey.

The rhythm becomes:

ARRIVE
  ↓
EXPLORE
  ↓
DISCOVER
  ↓
CONFIGURE
  ↓
DEPART
  ↓
PLATFORMING ROUTE
  ↓
ARRIVE AT NEXT PLANET

This makes reaching another planet feel like an actual journey rather than completing an arbitrary numbered stage.

The Three-Part Capability System

Each section of the campaign expands three major axes of the player's capability space.

1. Proto: Biology

Proto is the primordial side of ALife.

Proto discovers animals and can copy their morphology, instincts, or biological properties.

Examples:

Proto is particularly important for cleansing infection.

Different animals can provide fundamentally different ways of removing corruption from robots, machinery, or environments.

Proto answers:

What biological capability can I use?

2. Kee: Gadgets

Kee represents the robotic and mechanical side.

Kee can become or provide physical gadgets such as:

• sword,
• shield,
• bow,
• jetpack,
• grappling system,
• scanner,
• launcher,
• hover system,
• drill,
• drone,
• energy projector,
• and other mechanical forms.

Kee's gadget answers:

What physical/mechanical interface can I use?

A gadget and an algorithm are deliberately separate concepts.

The gadget is the body/tool.

The algorithm is the logic.

3. Kee: Real Computer-Science Algorithms

Algorithms in ALife are intended to be based on actual computer-science algorithms, not merely fictional abilities with technical names.

The broader design has a large library of possible algorithms to draw from.

Examples include:

• A*
• Dijkstra's algorithm
• Breadth-First Search
• Depth-First Search
• Flood Fill
• sorting algorithms
• graph algorithms
• scheduling algorithms
• optimisation algorithms
• pathfinding algorithms

Algorithms become gameplay logic.

Examples:

A*

Find or expose an optimal path through a traversal problem.

Breadth-First Search

Propagate an effect outward through neighbouring nodes in layers.

Dijkstra

Find a route according to minimum cost rather than simply minimum distance.

Flood Fill

Propagate cleansing through a connected infected region.

Sorting / prioritisation

Dynamically reorder targets according to distance, danger, corruption, value, or another property.

The algorithm answers:

How is the action organised, reasoned about, or applied?

The Combination

The fundamental build grammar is:

Proto Animal / Power
        +
Kee Gadget
        +
Algorithm
        +
Attributes
        +
Legendaries
        =
Player Build

This allows the same animal, gadget, or algorithm to behave very differently depending on what it is paired with.

The long-term objective is not simply to become numerically stronger.

It is to accumulate an increasingly large vocabulary of possible behaviours.

Campaign Progression

Campaign content introduces capabilities before the player permanently masters every possible combination.

A campaign section can introduce:

1. a new animal,
2. a new Proto cleansing capability,
3. a new Kee gadget,
4. a computer-science algorithm,
5. and a platforming challenge designed around their interaction.

The player experiences the intended combination in authored content.

Completing that content establishes that the capability has been discovered.

For algorithms in particular, discovery and permanent endgame ownership are intentionally different stages.

Algorithm Discovery vs Mastery

An algorithm can first appear as part of its campaign content.

For example:

Discover A*
    ↓
Use A* during its authored challenge
    ↓
Complete the challenge
    ↓
A* becomes eligible for that multiverse's endgame loot

The campaign therefore says:

I discovered and experienced this algorithm.

The endgame says:

I mastered it and can now build around it permanently.

This creates a direct connection between authored campaign content and the later loot chase.

Gear Philosophy

ALife does not require traditional character levelling.

There does not need to be a central progression of:

Level 1 → Level 20 → Level 50 → Level 100

Instead:

Small capability library
        ↓
More animals
        ↓
More gadgets
        ↓
More algorithms
        ↓
More attributes
        ↓
More legendaries
        ↓
More combinations
        ↓
Massive build space

Progression is primarily horizontal.

The artificial life becomes more versatile and expressive rather than simply acquiring an ever-larger number next to its name.

Normal Gear and Attributes

Normal campaign content can produce random attribute rewards.

Attributes are the tuning layer.

Possible examples include:

• movement speed,
• cleansing radius,
• cleansing strength,
• energy efficiency,
• cooldown reduction,
• acceleration,
• air control,
• gadget recharge,
• transformation speed,
• attack speed,
• defence,
• resource collection.

The intention is to avoid turning ALife into inventory landfill.

Instead of forcing the player to retain hundreds of nearly identical disposable items, useful attributes can become part of the player's persistent build library.

Conceptually:

Find attribute
    ↓
Unlock attribute
    ↓
Use it when constructing builds

Players then save configurations they actually care about.

Algorithm Sets

Permanent portable algorithms are represented by three-piece sets.

Example:

A* Piece I
A* Piece II
A* Piece III
      ↓
Complete A* Set
      ↓
A* becomes available as part of the build

The three-piece structure is intentional.

An algorithm should matter enough to occupy meaningful build space without consuming the entire equipment configuration.

This leaves room for legendary items and attribute customisation.

The core principle is:

Algorithm sets give you the system.

Legendary Gear

Legendary gear has a different purpose.

It should not simply duplicate the role of algorithm sets or exist only as a larger attribute number.

Legendary items are rule benders.

They alter how an existing capability behaves.

Examples:

• A* can recalculate while airborne.
• Cleansing one target automatically marks another.
• Using a Kee gadget charges a Proto ability.
• Using a Proto ability resets a Kee gadget.
• An algorithm evaluates multiple destinations instead of one.
• A traversal action gains an unusual interaction with another unlocked system.

The principle is:

Algorithm sets give you the system. Legendaries mutate the system. Attributes tune the system.

This separation gives each reward category a clear reason to exist.

Saved Builds

The important persistent object is not necessarily a backpack full of equipment.

It is the build.

For example:

build: Rift Farming
proto:
  animal: Falcon
  power: Wind Purge

kee:
  gadget: Jetpack

algorithm:
  set: A*

legendaries:
  - Momentum Loop
  - Chain Cleanse

attributes:
  - Air Control
  - Energy Efficiency
  - Movement Speed

Another saved build might be designed for:

• bosses,
• traversal,
• cleansing,
• multiplayer,
• speed,
• resource farming,
• Greater Rifts,
• or a particular Autopilot route.

Progress therefore becomes a collection of solutions rather than a warehouse of junk.

Multiverse Campaign Completion

The campaign ultimately leads across the planetary system to the final planet.

The last traversal should be a culmination of the multiverse's mechanics.

Planet 7 contains the major boss or source of infection.

Conceptually:

Planet 1
  ↓
Discover / cleanse / traverse
  ↓
Planet 2
  ↓
Discover / cleanse / traverse
  ↓
...
  ↓
Planet 6
  ↓
Final traversal
  ↓
Planet 7
  ↓
BOSS
  ↓
MULTIVERSE CLEANSED

The final world can potentially be visible much earlier in the campaign, giving the player a persistent destination to work toward.

As the earlier planets are cleansed, the state of the multiverse can visibly improve.

Endgame Unlock

Defeating the multiverse boss changes the role of that entire multiverse.

It is no longer only campaign content.

It becomes an endgame ecosystem.

Completion unlocks:

• Rifts,
• Greater Rifts,
• algorithm-set farming,
• legendary farming,
• rarer attributes,
• build experimentation,
• recording,
• Autopilot farming of mastered content.

So every completed multiverse remains relevant after its story is finished.

Rifts

Rifts procedurally remix the capabilities and environment of the multiverse.

They can reuse and recombine:

• planetary geometry,
• radial gravity,
• animals,
• Proto powers,
• Kee gadgets,
• algorithms,
• infection mechanics,
• enemies,
• hazards,
• and loot families.

Importantly, because the campaign itself is built around travelling between planets, procedural Rifts can also generate increasingly unusual planet-bridging routes.

The authored campaign teaches the player how the multiverse works.

Rifts remix what was learned.

Greater Rifts

Greater Rifts are the scalable endgame frontier.

Players use them to:

• test builds,
• hunt algorithm set pieces,
• hunt legendaries,
• find better attributes,
• optimise combinations,
• push increasingly difficult content,
• record successful strategies,
• and eventually automate content they have mastered.

The game therefore retains an action-RPG farming loop without requiring traditional XP levelling.

Why Old Multiverses Stay Relevant

Each multiverse has its own discoveries and endgame loot pool.

For example, if an earlier multiverse contains the A* algorithm set, a player may return much later because A* interacts beautifully with an animal or gadget discovered several multiverses afterward.

New content therefore expands the usefulness of old content.

The ideal result is:

Every new multiverse increases the number of meaningful combinations across the entire game.

Recording and Autopilot

ALife includes a record-and-replay system for mastered content.

The player first performs the activity manually.

The system can record:

• actions,
• traversal,
• ability use,
• gadget use,
• transformations,
• interactions,
• and the build used for the run.

That creates a reusable behavioural/data cache.

Conceptually:

Player discovers challenge
        ↓
Player solves challenge manually
        ↓
Player records successful solution
        ↓
ALife replays learned solution
        ↓
Solved content can be farmed

The long-term goal is not merely a dumb input macro.

ALife is intended to support artificial-life behaviour that can run and learn at the same time, allowing mastered routines to potentially improve or adapt.

The fundamental rule is:

The player pushes the frontier. ALife farms what has been mastered.

Full Gameplay Loop

ENTER MULTIVERSE
       ↓
ARRIVE ON PLANET
       ↓
DISCOVER ANIMAL / PROTO POWER
       ↓
DISCOVER KEE GADGET
       ↓
EXPERIENCE ALGORITHM
       ↓
BUILD A COMBINATION
       ↓
DEPART PLANET
       ↓
3D RADIAL PLANET-BRIDGING PLATFORMING
       ↓
CLEANSE INFECTION
       ↓
ARRIVE AT NEXT PLANET
       ↓
REPEAT ACROSS THE SYSTEM
       ↓
FINAL PLANET
       ↓
BOSS
       ↓
MULTIVERSE CLEANSED
       ↓
RIFTS + GREATER RIFTS UNLOCK
       ↓
FARM 3-PIECE ALGORITHM SETS
       ↓
FARM LEGENDARIES
       ↓
TUNE ATTRIBUTES
       ↓
SAVE BUILDS
       ↓
PUSH HARDER RIFTS
       ↓
RECORD MASTERED ROUTES
       ↓
AUTOPILOT FARMING
       ↓
ENTER THE NEXT MULTIVERSE

Progression in One Sentence

Campaign content teaches capabilities, algorithm sets make algorithms permanently portable, attributes tune the build, and legendaries break its rules.

World Structure in One Sentence

The seven planets are the chapters; the journeys between them are the levels; the final planet is the boss destination; cleansing the system turns the whole multiverse into an endgame playground.

ALife's Design Identity

ALife is ultimately about accumulating possibility.

Proto discovers nature.

Kee discovers machinery and computation.

The player combines them.

The campaign teaches those systems through compact radial-gravity platforming journeys between planets.

The endgame takes the systems the player has learned and explodes them into buildcraft, procedural Rifts, Greater Rifts, legendary modifiers, algorithm sets, recorded solutions, and autonomous farming.

There is no requirement for the player to become "Level 100."

Instead, the artificial life becomes something much stranger:

A growing library of forms, tools, algorithms, behaviours, builds, and learned solutions that belongs to the same persistent being across every multiverse.