Opus 5.5 is good at explainer videos
Launchvideo uses Claude Opus 5.5 in a serverless agent to render explainer MP4s from a URL or prompt.
Launchvideo.io turns a URL or prompt into a product explainer by having Claude Opus 5.5 write the film as code, which a serverless OpenComputer agent renders to MP4. Each job runs in a fresh Amazon Linux 2023 arm64 microVM with 4 vCPU and 8 GB RAM, using roughly 90k input and 15k output tokens and finishing in about four minutes. Three tools fetch the site, check scenes for errors and visible text, and pipe 1920x1080 JPEG frames at 30 fps through ffmpeg. The agent holds no secrets; a Vercel Blob upload token scoped to one path lasts three hours.
- Opus 5.5 writes each explainer as code, not with a video model.
- A typical film uses about 90k input and 15k output tokens.
- Rendering uses headless Chromium and ffmpeg in a disposable microVM.
- The TypeScript agent deploys to OpenComputer with one command.
Full article487 words · extracted from launchvideo.io · click to collapse
Paste a URL or describe the product. Opus 5.5 writes the film and a serverless agent renders it. About four minutes and roughly 100k tokens per video.
examples
Made by this page, untouched.
Each one is a single run: a URL or a prompt in, an MP4 out. No edits.
NVIDIA
nvidia.com
Jev, by TypeSafe AI
typesafe.ai
OpenComputer
opencomputer.dev
Linear
linear.app
Infera (from a prompt)
“make a modern slick and punchy video for a modern startup that works on inference”
how it runs
A serverless agent on OpenComputer. Yours in one click.
The whole product is one agent file, three tools, and this form. OpenComputer runs the agent, the microVM it renders in, the model gateway, and the session API the page polls.
- One OpenComputer serverless agent, defined in TypeScript and deployed with
opencomputer deploy. No framework, no queue, no server of ours. - anthropic/claude-opus-5.5 through OpenComputer's model gateway. Roughly 90k input and 15k output tokens per film, most of it the HTML itself.
- Every job is one session in a fresh microVM: Amazon Linux 2023 on arm64, 4 vCPU, 8 GB RAM, Node 22. The first tool call installs Playwright's headless Chromium and a static ffmpeg (about a minute); the VM is thrown away after.
- Three
defineToolfunctions. web_fetch returns page text plus title, headings, the most used hex colors, and Google Fonts. check_scene loads the film and reports JS errors and the visible text at sample timestamps. render_video renders and uploads. - No video model. The page's clocks (requestAnimationFrame, timers, Date, CSS and Web Animations) are replaced with a virtual clock, so every frame is a deterministic seek. 1920x1080 at 30 fps, JPEG frames piped into libx264, crf 18.
- The agent holds no secrets. The form mints a Vercel Blob upload token scoped to one path for three hours, parks it in a per-job manifest, and the tool fetches it by job id. The finished MP4 is a public Blob URL.
- This page uses the same API the CLI does: create a session, send one turn, poll the event stream (tool.started, tool.completed, turn.completed) to show progress, and treat the MP4 appearing in Blob as done.
Agent
Model
Runtime
Tools
Rendering
Storage
Control plane
// opencomputer/agents/director/agent.ts
import { useInput, useModel, useTool } from "@opencomputer/agent";
import { checkScene, renderVideo } from "./tools/scene.js";
import { webFetch } from "./tools/web.js";
export default function Agent() {
const input = useInput(); // the JOB block from the form
useModel("anthropic/claude-opus-5.5");
useTool(webFetch); // read the product's site
useTool(checkScene); // load the HTML, report errors + visible text
useTool(renderVideo); // headless Chromium → ffmpeg → Blob
return `You are a motion designer who writes code. ...`;
}$npx opencomputer template deploy https://github.com/diggerhq/shipvideo
Everything above is in the repo, and one click deploys it to your account. The idea comes from Deedy's post on Opus 5.5 and instructional video: the model writes the film as code, and code renders the same every time.