Skip to content

Built for the work.
Not for the feed.

WebGL.com and WebGPU.com are being rebuilt as one home for GPU work: a showcase filed under an opinionated taxonomy, credited to the people who made it, and searchable by the techniques behind it. Anyone may submit. What gets seen is earned.

Open to submissionsSubmit a project (opens in a new tab)

Anyone may submit. A person reads every one.

Connected foundations, not competing silos.

WebGL and WebGPU are the connected foundations of a cross-platform GPU community. WebGL established portable real-time graphics for the web and remains essential. WebGPU advances that model with modern graphics and GPU computing for the web, while implementations such as Dawn and wgpu extend the wider ecosystem into native and other non-browser environments.

We do not treat WebGPU as browser-only, WebGL as obsolete, or the two as competing silos. We cover the projects, people, tools, and knowledge moving real-time graphics and GPU computation forward across platforms, while making runtime-specific claims only when the evidence supports them.

What we're building toward.

Real-time graphics has never had a home that treated it as craft. The work scatters across social feeds that compress it, autoplay it muted, and bury it in a day. Those feeds were never built to render a shader properly, let alone to let you find one again six months later. The archive that survives is a pile of dead links.

This is a rebuild of both domains around the opposite premise: that a piece of GPU work is worth filing, crediting, and searching by how it was made. What follows is what we are building toward and what we will not trade away, written down plainly so it can be held against us later.

Finding the work is the hard part.
The single biggest complaint in this field is that good work is scattered across a dozen platforms and none of them are searchable. One place, faceted, and built so a technique you saw once is findable on purpose rather than by luck.
Visibility is earned, and never gated.
Merit is the baseline and it is free: great work surfaces because it is good. Paid options may add reach on top of that baseline and are always labeled, but nothing paid displaces or outranks unpaid merit, and being seen is never something you buy your way past.
Credit is structural, not a courtesy.
Every piece names the person who made it, and that link is part of the data model rather than a caption. Work that travels should carry its author with it.
A quality floor beats an open flood.
Anyone may submit; not everything earns the same room. In an era of infinite generated output, a curated shelf is worth more than a complete one, so we grow around quality, with a way in for anyone willing to earn it.
How it was made is the interesting part.
The taxonomy is opinionated on purpose: it files work by technique, not just by subject, because "how did they do that" is the question that actually brings people here.
Small on purpose.
This is built to be a healthy small thing that lasts, not a land grab with an exit. That is a constraint we chose, and it decides what we build and what we turn down.

Send us the work.

Submissions run through one form, and a person reads every one. Tell us what you built and where it runs; we file it under the taxonomy, credit you by name, and point back at wherever you want people to follow you.

The form is for projects: the thing you made, running somewhere we can open it. Guides, tools, news and corrections have no intake of their own yet, so they go to the inbox below and get handled by hand.

Submit a project (opens in a new tab)

Opens a form in a new tab. It asks what you built, where it runs, and how you want to be credited.

What a good submission has
Say what it is, and who it is for.
The description is the field that decides whether the entry reads well: what the thing does, who it is for, and what is interesting about how it was built. A few sentences beat a tagline.
Link something we can open.
A live demo is the strongest submission; a video carries the work that cannot be hosted. The repo field is optional and worth filling anyway. Source is the difference between “nice” and “now I can learn from it”.
Name the techniques in the description.
There is no tag picker on the form, so write them down: compute passes, raymarching, instancing, whatever the work actually leans on. We file by technique, and that is how somebody finds this again six months from now.
Credit everyone who touched it.
Collaborators, the engine or library it stands on, anything you ported or learned from. Credit is part of the record here rather than a line at the bottom, so tell us who else belongs on it.
Say what you would be up for.
The last question asks whether you would do a walkthrough, a Q&A, a write-up, or an introduction to people working on the same problems. It is the fastest way for a submission to become more than a listing.

One inbox, read by a human.

Corrections, submissions, security disclosures, or anything else all go to one inbox. There is no routing maze and no contact form. Mail goes to a person, and you will get an answer even when the answer is no.

admin@webgpu.com