Archetype ECS Lib

A tiny archetype-based ECS (Entity Component System) for TypeScript.

• Archetypes (tables) store entities in a SoA layout (one column per component type).
• Queries iterate matching archetypes efficiently.
• Commands let you defer structural changes (spawn/despawn/add/remove) safely.
• A minimal Schedule runs systems by phases and flushes commands between phases.

Exports are defined in index.ts:

• Types, TypeRegistry, Commands, World, Schedule

Install

npm i archetype-ecs-lib
# or
pnpm add archetype-ecs-lib
# or
yarn add archetype-ecs-lib

Core concepts

Entity

An entity is a lightweight handle:

type Entity = { id: number; gen: number };

The gen (generation) prevents using stale entity handles after despawn/reuse.

Component

A component is any class used as a type key:

class Position { constructor(public x = 0, public y = 0) {} }
class Velocity { constructor(public x = 0, public y = 0) {} }

Internally, constructors are mapped to a stable numeric TypeId via typeId().

World

World owns entities, archetypes, commands, and systems.

Structural operations:

• spawn(), despawn(e)
• add(e, Ctor, value), remove(e, Ctor)
• has(e, Ctor), get(e, Ctor), set(e, Ctor, value)
• query(...ctors) to iterate entities with required components
• cmd() to enqueue deferred commands
• flush() applies queued commands
• update(dt) runs registered systems and flushes at the end

Deferred structural changes (important)

While iterating a query (or while systems are running), doing structural changes directly can throw:

“Cannot do structural change (…) while iterating. Use world.cmd() and flush …”

Use world.cmd() inside systems / loops, and let world.flush() apply changes safely.

Quick start

import { World, Schedule } from "archetype-ecs-lib";

class Position { constructor(public x = 0, public y = 0) {} }
class Velocity { constructor(public x = 0, public y = 0) {} }

const world = new World();

// Spawn immediately
const e = world.spawn();
world.add(e, Position, new Position(0, 0));
world.add(e, Velocity, new Velocity(1, 0));

// A simple system
world.addSystem((w: any, dt: number) => {
  for (const { e, c1: pos, c2: vel } of w.query(Position, Velocity)) {
    pos.x += vel.x * dt;
    pos.y += vel.y * dt;

    // Defer structural changes safely
    if (pos.x > 10) w.cmd().despawn(e);
  }
});

world.update(1 / 60);

Note: SystemFn is typed as (world: WorldI, dt) => void where WorldI only requires flush(). In practice, you’ll typically use the concrete World API in systems (cast world or type your function accordingly).

Query API

for (const row of world.query(Position, Velocity)) {
  // row.e  -> Entity
  // row.c1 -> Position
  // row.c2 -> Velocity
}

query(...ctors) yields objects shaped like:

• e: the entity
• c1, c2, c3, …: component values in the same order as the ctor arguments

So if you call query(A, B, C) you’ll get { e, c1: A, c2: B, c3: C }.

Commands API (deferred ops)

const cmd = world.cmd();

cmd.spawn((e) => {
  cmd.add(e, Position, new Position(0, 0));
});

cmd.add(entity, Velocity, new Velocity(1, 0));
cmd.remove(entity, Velocity);
cmd.despawn(entity);

// Apply them (World.update() also flushes automatically at end)
world.flush();

Supported commands (Commands.ts):

• spawn(init?)
• despawn(e)
• add(e, ctor, value)
• remove(e, ctor)

Schedule (phases)

Schedule is a small phase runner:

import { World, Schedule } from "archetype-ecs-lib";

const world = new World();
const sched = new Schedule();

sched
  .add("input", (w: any) => { /* read input, enqueue commands */ })
  .add("sim",   (w: any, dt) => { /* update movement */ })
  .add("render",(w: any) => { /* build render data */ });

const phases = ["input", "sim", "render"];

// Runs each phase in order and calls world.flush() after each phase.
sched.run(world, 1/60, phases);

This is handy when you want deterministic ordering and command application points.

World API summary

Entity lifecycle

• spawn(): Entity
• despawn(e: Entity): void
• isAlive(e: Entity): boolean

Components

• has(e, Ctor): boolean
• get(e, Ctor): T | undefined
• set(e, Ctor, value): void (requires the component to exist; otherwise throws)
• add(e, Ctor, value): void (structural: may move entity between archetypes)
• remove(e, Ctor): void (structural: may move entity between archetypes)

Systems / frame

• addSystem(fn): this
• update(dt): void (runs systems in order, then flushes)
• cmd(): Commands
• flush(): void

Queries

• query(...ctors): Iterable<{ e: Entity; c1?: any; c2?: any; ... }>

Notes & limitations

• This is intentionally minimal: no parallelism, no borrow-checking, no automatic conflict detection.
• Query results use c1/c2/... fields for stability and speed; you can wrap this in helpers if you prefer tuple returns.
• TypeId assignment is process-local and based on constructor identity (WeakMap).

License

MIT License

Copyright (c) 2025 Jean-Laurent Duzant

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.