Repolode
diggerhq

diggerhq/shipvideo

Turns a pasted URL or a text prompt into a 20-40 second launch video, written as HTML by Claude Opus and rendered to MP4 by a serverless agent.

Stars
351
Forks
39
Open issues
1
Last commit
19 days ago

What is it?

LaunchVideo (shipvideo) generates short launch videos without a video model. Given a URL or a prompt, Claude Opus 5.5 writes a single HTML film, and a serverless agent on OpenComputer renders it frame by frame in headless Chromium under a virtual clock, then encodes the frames with ffmpeg into an MP4.

In URL mode, the agent first fetches the page and pulls out the title, description, headings, dominant hex colors and Google Fonts, so the video reflects the product's own branding. The agent checks the scene for JS errors and visible text before rendering, and the finished file is saved to the session workspace for download.

The project is split into an OpenComputer agent project (the `director` agent) and a Next.js web app that provides the form and two API routes. The web app needs only an OpenComputer API key and keeps no storage of its own.

Key features

  • 01Generates a 20-40 second 1920x1080 video from a URL or a text prompt.
  • 02Renders frames in headless Chromium under a virtual clock for deterministic output.
  • 03Encodes the result to MP4 with ffmpeg (libx264, yuv420p, faststart).
  • 04Extracts colors, fonts and page text from a URL to match the product's branding.
  • 05Includes a one-click deploy manifest for OpenComputer accounts.

Use cases

  • Creating a quick product launch video from an existing landing page URL.
  • Producing short explainer clips from a text description without video editing software.
  • Experimenting with LLM-generated HTML animation rendered to video.

Quick start

  1. 1

    Install dependencies and link the OpenComputer project

    npm install && npx opencomputer link --create-project shipvideo && npx opencomputer deploy --watch
  2. 2

    Install web app dependencies and pull environment variables from Vercel

    cd web && npm install && vercel link && vercel env pull .env.local
  3. 3

    Add your OpenComputer API key and agent ID to web/.env.local

  4. 4

    Start the web app locally

    cd web && npm run dev

Things to know

  • Requires an OpenComputer account and API key to run.
  • The README does not state a license, so reuse terms are unclear.
  • Rendering takes roughly real time, so a 30 second film takes 30 to 40 seconds.

Topics

FAQ

What is shipvideo?
Turns a pasted URL or a text prompt into a 20-40 second launch video, written as HTML by Claude Opus and rendered to MP4 by a serverless agent. LaunchVideo (shipvideo) generates short launch videos without a video model. Given a URL or a prompt, Claude Opus 5.5 writes a single HTML film, and a serverless agent on OpenComputer renders it frame by frame in headless Chromium under a virtual clock, then encodes the frames with ffmpeg into an MP4.
Is shipvideo free to use? What license does it have?
shipvideo has no declared license. Its code is public, but without a license you have no explicit permission to reuse it; ask the author before relying on it.
Is shipvideo still maintained?
Actively maintained. The last commit to shipvideo was on Oct 1, 2026.
What can I use shipvideo for?
Creating a quick product launch video from an existing landing page URL. Producing short explainer clips from a text description without video editing software. Experimenting with LLM-generated HTML animation rendered to video.
How do I install shipvideo?
Install dependencies and link the OpenComputer project: npm install && npx opencomputer link --create-project shipvideo && npx opencomputer deploy --watch Install web app dependencies and pull environment variables from Vercel: cd web && npm install && vercel link && vercel env pull .env.local Add your OpenComputer API key and agent ID to web/.env.local Start the web app locally: cd web && npm run dev
Where can I get shipvideo?
The source code and installation instructions are on GitHub: https://github.com/diggerhq/shipvideo

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects