@thefehr/foundry-playwright

A robust, multi-version E2E testing library for FoundryVTT modules and systems, powered by Playwright.

Core Features

• Multi-Version Support: Built-in adapters for FoundryVTT V13 and V14 with automatic version detection.
• Backup-Based World Reset: useBaseWorld snapshots the world state before the first spec and restores it before each test — fast, isolated, no manual teardown (V14). Falls back to full recreate on V13.
• Docker Orchestration: Automated setup and teardown of version-specific Foundry instances via DockerFoundryOrchestrator or the foundry-playwright test CLI.
• State Manipulation: Fast, UI-less data injection via direct Foundry API and socket calls (createDocument, createTestActor, grantCurrency, createUser, setRolePermission, and more).
• UI Helpers: Tab switching, drop simulation, aggressive tour suppression, and dialog automation across V13 and V14 sheet layouts.
• System Adapters: First-class support for dnd5e and PF2e with pluggable SystemStateAdapter and UIAdapter interfaces.
• Verified Versions Matrix: A continuously updated verified-versions.json registry of confirmed-working (Foundry × system) combinations, maintained by a nightly release monitor. Ships with the published package and readable via getVerificationRegistry() from consumer code, e.g. to diff a downstream project's own CI matrix against it.
• Screen vs. Viewport: withScreenSize/setScreenSize let window.screen.width/height (the physical display, used by some modules for mobile detection) diverge from viewport (window.innerWidth/height), which Playwright's viewport option alone can't do.

Getting Started

npm install --save-dev @thefehr/foundry-playwright

npx foundry-playwright init

init scaffolds a playwright.config.ts, an e2e/ directory with a sample test, and a test:e2e script in your package.json.

Writing Tests

useFoundry — simple setup/teardown

Creates the world in beforeAll, tears it down in afterAll. Good for quick, stateless tests.

import { test, expect, useFoundry } from "@thefehr/foundry-playwright";

useFoundry(test, {
  worldId: "test-world",
  systemId: "dnd5e",
  moduleId: "my-module-id",
});

test("actor can be created", async ({ foundry }) => {
  await foundry.state.createTestActor("Hero");
});

useBaseWorld — backup-based isolation (recommended for V14)

Takes a one-time snapshot after setup, then restores it before every spec. Each test starts from a clean slate without paying full setup cost.

import { test, useBaseWorld, verifyResult } from "@thefehr/foundry-playwright";

useBaseWorld(test, {
  worldId: "my-test-world",
  systemId: "dnd5e",
  moduleId: "my-module-id",
  setupWorld: async ({ state }) => {
    await state.createTestActor("Base Actor");
  },
});

test("currency grant is logged", async ({ page, foundry }) => {
  await foundry.state.grantCurrency("Base Actor", 100, "gp");
  await verifyResult(page, "currency-update", (d) => d.amount === 100);
});

Running Tests

# Against a running Foundry instance
npx playwright test

# With a Docker-managed Foundry instance
npm run verify:local -- --docker --version 14.360.0 --system dnd5e

# foundry-playwright test --docker auto-mounts e2e/<name>/module.json folders and a
# root module.json in the CWD. For a module that needs a build step first (e.g. a
# git-submodule checkout built into its own dist/), point at the built output directly:
npx foundry-playwright test --docker --module-dir ../my-module/dist

Breaking change: foundry-playwright test --docker no longer reads a .env file from your project directory on its own. Export FOUNDRY_USERNAME, FOUNDRY_PASSWORD, and FOUNDRY_ADMIN_KEY as real environment variables before running the CLI (or load your own dotenv first). This also removes the envFile option from DockerOrchestratorConfig for anyone using DockerFoundryOrchestrator programmatically — pass username/password/adminKey directly instead.

Configuration

| Variable | Purpose | | :--------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------------------- | | FOUNDRY_URL | Base URL of the Foundry instance (default: http://localhost:30000) | | FOUNDRY_VERSION | Force V13 or V14 adapter instead of auto-detecting | | FOUNDRY_SYSTEM_ID | Active game system (default: dnd5e) | | FOUNDRY_UI_ADAPTER | UI adapter: default, dnd5e, or tidy5e | | FOUNDRY_ADMIN_PASSWORD / FOUNDRY_ADMIN_KEY | Admin password for setup operations | | FOUNDRY_USERNAME / FOUNDRY_PASSWORD | Foundry account credentials for Docker image download | | FOUNDRY_PLAYWRIGHT_ROOTLESS | Set to 1 when --docker runs against rootless Podman, so bind-mounted directories keep correct ownership (see DockerOrchestratorConfig.rootless) |

UIAdapter (and registerUIAdapter) is scoped to actor-sheet-shaped UIs — an application selector plus tabs/collapsible sections. If the module under test has its own UI that isn't an actor sheet variant (a HUD, a control bar, a custom fullscreen layout), define plain selector constants in your own test suite instead of forcing that UI through the adapter interface.

Documentation

• Authentication & World Setup
• State Manipulation Fixtures
• Canvas Interaction
• System Agnosticism & Adapters
• Multi-Version Support
• Docker Orchestrator
• Getting Started / Migration Guide

Contributing

See CONTRIBUTING.md for setup, local checks, and PR conventions. This project follows a Code of Conduct. Found a security issue? See SECURITY.md instead of opening a public issue.