PhaserForge

Project docs:

Design the scene. Compose the behavior. Tune it live.

PhaserForge is a gameplay behavior editor for Phaser built around one bet: the real slowdown is not movement code, collision code, or rendering code. It is orchestration.

The drag comes from timers, flags, ad-hoc state machines, event-order bugs, and one more special case buried in update(). PhaserForge pushes that glue into an ACE model: author Actions, gate them with Conditions, and trigger them from Events.

The claim is narrow on purpose. This is not “a better everything-editor.” It is a proof-of-concept that if orchestration becomes declarative, inspectable, and live-tunable, then a real kind of 10x game development becomes possible for gameplay-heavy teams.

Why It’s Different

ACE authoring	Live tuning	Declarative reuse
Build behavior as events + actions + conditions, not glue code.	Change values while the game is running and feel the result immediately.	Keep gameplay logic in YAML so it can be inspected, versioned, duplicated, and remixed.

Highlights

Design	Orchestrate	Ship
Multi-scene projects, base-scene-plus-waves, formations, text entities, scene graph editing, layout tools, and drag/drop asset workflows.	Event blocks, loops, patterns, parallel actions, counters, collections, collisions, triggers, semantic input maps, and live preview.	YAML round-trip, project library, online/offline workspace flows, cloud account support, GitHub connect, and GitHub Pages publishing.

Less glue. More game.

Start Here

• DeepWiki for architecture and feature details
• Getting Started for the guided path
• Workflow Reference for exact controls and editor flows
• .repo-memory/product-memory.md for compact repo-wide invariants and .repo-memory/regression-playbook.md for regression-prevention habits

Requirements

Node.js is required to install dependencies and run scripts via npm.

Commands

Command	Description
npm install	Install project dependencies
npm run dev	Start the dev server (defaults to http://localhost:8080)
npm run dev:cloud	Start the dev server + local API (Vite proxies /api/* to the API)
npm run build	Create a production build in dist/
npm test	Run unit tests (Vitest)
npm run test:stories	Run Storybook-focused tests
npm run test:e2e	Run Playwright end-to-end tests
npm run test:all	Run unit + e2e tests
npm run docs:dev	Start the docs site locally

For deeper setup, workflows, testing conventions, and cloud publishing details, use the docs and DeepWiki instead of treating this README as the full manual.

Lightweight Repo Memory

This repo keeps durable, high-value product memory in .repo-memory/:

• product-memory.md: short product-wide invariants that future changes should preserve
• regression-playbook.md: compact guidance for turning repeated bug classes into lasting guardrails

Everything else stays where it already fits best:

• tests for executable guarantees
• top-level .plans/ for the small active workflow/proposal surface
• .plans/archive/ for historical plans and mockups that should not drive new implementation by default
• scoped AGENTS.md files for local rules near the code they govern

About log.js

npm run dev / npm run build run node log.js ... in the background, which sends a lightweight anonymous GET request (event + Phaser version + package name). If you’d rather not send this, use npm run dev-nolog / npm run build-nolog.