Coding

Claude Code Built a 61 KB WebGPU Ocean Sim in About 30 Minutes From 1 Prompt

  • Level Advanced
  • Setup1–3 hrs (estimated)
  • Evidence Some evidence
  • CostNot stated in source
  • No-codeNo
Claude CodeSource-linked

What they did

A Reddit user gave Claude Code one prompt: make the most realistic water scene it could in WebGPU, with a floating camera. It wrote the renderer from scratch with no engine or libraries (FFT waves, whitecaps, a physically based sky and sun glitter) in a single 61 KB HTML file. It screenshotted its own output and had separate subagents critique the screenshots so it wasn't judging its own work. The whole thing took about 30 minutes.

How it went

A 61 KB HTML ocean you can fly around in a browser.

Worth knowing

  • Popular Reddit post (200+ comments); graphics veterans in the thread point out ocean shaders like this are a classic demo, so it shows the workflow more than a breakthrough. Needs a WebGPU browser; it ran slowly on a phone.

Prompt

Original prompt, as posted by the author.

Make the most realistic [ocean / water] scene you can in WebGPU, in a single HTML file with no engines or libraries, and a free-flying camera (WASD + mouse). After each version, take screenshots and have a separate subagent critique them for realism; fix the top issues and repeat until the critiques stop finding big problems.

Recommended

Try it in Grok Bot

Originally done with Claude Code; this prompt works in Grok Bot.

  1. Copy the prompt (button below).
  2. Open Grok Bot chat and paste it. Replace any [bracketed] placeholders.
  3. Connect the listed tools (Reddit) as connectors when Grok Bot asks.
  4. If the task repeats, ask: "Make this a daily routine."

Alternatives: the same prompt works in Claude Code, Claude, ChatGPT Agent, OpenClaw.

Set it up

Tools: Reddit · Stack: Claude Code (Opus 5.5)

  1. Read the original source for full context.
  2. Connect the tools your agent needs.
  3. Paste the prompt and replace [bracketed] placeholders.
  4. Run once manually and check the output before scheduling it.

Generic suggested steps, not from the source.

Evidence

Popular Reddit post (200+ comments); graphics veterans in the thread point out ocean shaders like this are a classic demo, so it shows the workflow more than a breakthrough. Needs a WebGPU browser; it ran slowly on a phone

Source: Reddit · 2026-10-16 · Verified by us: not yet · Report broken or outdated

Related use cases

ContentX · 2026-01

Turn YouTube channels into a weekly EPUB magazine

Every week, pulls new videos from a list of YouTube channels, rewrites their transcripts as magazine articles, and emails them to you as a newsletter plus an EPUB file.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Strong evidence
Claude CodeSource-linked
CodingX · 2026-10

Hermes Agent Nearly Doubled Speech-to-Text Accuracy on Hard Words: 99 to 183 of 240

A user turned his 15K saved Wispr Flow dictations into a test set for the local Whisper model behind his Hermes Agent. The agent turned his Wispr dictionary into a hint list, diffed raw vs corrected text to find words Whisper keeps mishearing, made find-and-fix rules, and tested every rule against the full history. With no fine-tuning, dictionary words went from 99/240 to 183/240 on Wispr clips and 19/45 to 43/45 on sentences he read aloud. It now archives the data nightly and looks for new fixes weekly.

  • Level Advanced
  • Setup1–3 hrs*
  • Evidence Strong evidence
Hermes AgentSource-linked
ResearchX · 2026-03

Build a NotebookLM + Obsidian research vault

Claude Code sends a pile of videos, PDFs and articles through NotebookLM and turns the output into linked Obsidian notes covering scenarios, actors and probabilities.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Some evidence
Claude CodeSource-linked