Three.js WebGPU Skill Pack

Guides

Three.js WebGPU skill-pack guides, organized by the decision you need to make

Decision and adoption guides for choosing, installing, migrating, and verifying the Three.js WebGPU skill pack with repository-backed evidence.

Start here when you know the decision you need to make but not the page you need. Choose a path for installation, skill routing, migration, implementation, agent use, or troubleshooting.

Published 2026-07-16Reviewed 2026-07-16

Start with the decision, not the folder name

The skill pack covers rendering architecture, procedural systems, final-image work, and validation. That breadth is useful only when the first decision is clear. This hub separates adoption questions from implementation questions so that a developer or coding agent can load the smallest relevant context.

If you are still deciding whether the pack fits your work, start with who the pack is for. Graphics engineers can follow the rendering architecture path. Technical artists can follow the materials and procedural-art path. Existing Three.js developers can use the WebGPU adoption path. Teams operating coding agents can begin with the agent-user path.

Install and use the pack

Use the documentation hub for operating instructions rather than product comparisons.

Agent-facing contracts have their own surface. Read the agent documentation hub for the division between project ownership and skill ownership, then use routing and minimal context to keep the active context narrow.

Migrate an existing workflow

Migration pages own step-by-step transition queries. The migration hub explains how the individual paths fit together.

These guides do not assume that changing an import completes a migration. Renderer initialization, material semantics, pass ownership, output conversion, readback, and evidence all need explicit checks against the installed Three.js revision.

Compare approaches before committing

Use the comparison hub when the decision involves two named approaches.

If the question is broader than one named competitor, use the alternatives hub, alternatives to Three.js agent skills, or alternative Three.js WebGPU learning resources.

Apply the pack to a concrete workload

Industry pages are workload maps, not decorative persona pages. Start with industry workflows, then choose browser-game rendering or product visualization and configurators. Each page states what the graphics skills can own and which application concerns remain outside the pack.

The pack itself costs nothing to license. The pricing page separates that fact from the real costs of engineering time, target devices, assets, tooling, and hosted services.

Resolve a concrete question

The FAQ hub is an index of questions observed in real documentation, repository, and community work. Each answer has its own canonical route.

What this hub does not claim

This page is navigation, not proof that every Three.js workload is covered. The installed package, current Three.js source, project requirements, and evidence bundles whose recorded source hashes still match remain authoritative. A lab can establish a mechanism under its declared capture contract, but it cannot establish performance on an unnamed device or success in an unrelated production scene.

The installed Choose Skills contract defines the routing inputs, causal-owner selection, cross-system handoffs, output ownership, explicit gaps, and verification result. The native WebGPU validation harness documents the intended rendering-evidence structure, but its checked summary must be regenerated when its source hash changes. Neither contract is a substitute for running the relevant checks in the project being changed.

Relevant product evidence

Evidence from webgpu-validation-harness relevant to Three.js WebGPU skill-pack guides, organized by the decision you need to make
Source lab: webgpu-validation-harness. Canonical lab; current status accepted. Inspect the evidence report.

Next steps

Relevant skills

Relevant demos and evidence

Sources and correction path