AI/TLDR

html-video

Your coding agent writes the HTML; your laptop renders the MP4

Video GenerationOpen source
Language
HTML
License
Apache-2.0
$npx playwright install chromium

Overview

html-video is a local studio and CLI for making short videos with a coding agent instead of a timeline editor. You describe a video — or paste an article link or a GitHub repo — and the agent picks a template, fills it with your content across multiple animated frames, and the tool records the result into a real MP4 on your own machine. The whole render happens locally: headless Chromium captures the animated HTML frame by frame and ffmpeg encodes it with libx264, so there is no cloud render step and no per-clip fee.

The project positions itself as a meta-layer over the HTML-to-video category rather than another authoring model to learn. Engines sit behind a single render(input, ctx) adapter contract; the Hyperframes engine (HTML + CSS + GSAP) is shipped and fully wired up today, while Remotion, Motion Canvas/Revideo and Manim are on the roadmap with their adapters not yet built. The README's engine table is explicit about which of those is actually runnable, and the maintainers treat it as the source of truth.

Around that sits the practical surface: 21 curated single-file HTML templates covering data visualisation, product promos, social shorts, explainers, kinetic type and transitions; a content graph that drives multi-scene storyboards with inline per-frame text editing; optional AI background music and narration mixed in at export; and auto-detection of 14 coding agents on your PATH, from Claude Code and Codex to Cursor, Gemini CLI, Aider and OpenCode. It is Apache-2.0 licensed with no seat caps or contributor agreements.

What it does

  • Local MP4 rendering — headless Chromium records the animated HTML and ffmpeg encodes it, with no cloud render and no per-render fee
  • Pluggable engine architecture behind one render adapter interface; the Hyperframes engine ships working today
  • 21 curated, license-clean animated templates: data viz, hero frames, cinematic looks, VFX text, logo outros and more
  • Article, repo or prompt as the input — the studio fetches a pasted URL server-side and builds the video from the real content
  • Multi-frame storyboards driven by a content graph, with inline per-frame text edits, reordering and re-render
  • 14 coding agents auto-detected on your PATH and switchable from the studio's top bar, plus a scriptable CLI
  • Optional AI soundtrack — background music and narration mixed into the MP4 at export

Getting started

You need Node.js 20+, pnpm 9+, a recent ffmpeg and a Chromium install. If you do not have a system Chromium, install Playwright's.

Install Chromium for the renderer

bashbash
npx playwright install chromium

Build and open the studio

The studio opens at http://127.0.0.1:3071. From there you pick a template or describe a video, chat with your agent, edit per-frame text, add a soundtrack and export MP4.

bashbash
pnpm install
pnpm -r build
node packages/cli/dist/bin.js studio

Check what the CLI can see

doctor reports which coding agents and engines are installed; search-templates finds a starting template for an intent.

bashbash
node packages/cli/dist/bin.js doctor
node packages/cli/dist/bin.js search-templates --intent "github stars race" --top 3

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

When to use it

  • Turn a blog post or release announcement into a short animated explainer without opening a video editor
  • Generate a data-viz clip from numbers you already have, using an editorial chart template the agent fills in
  • Produce many branded social cuts from one storyboard, re-rendering locally instead of paying per export
  • Script video generation from a repo's CI using the CLI rather than the studio UI

How html-video compares

html-video alongside other open-source video generation tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
MoneyPrinterTurbo★ 125kA self-hosted WebUI, API and CLI that turns one topic or keyword into a finished short video — script, stock or generated footage, voiceover, subtitles and music.
OpenMontage★ 60.4kAn agentic video production system that gives a coding assistant 12 pipelines and 100+ tools to research, script, generate and render a video end to end.
HyperFrames★ 51.9kHeyGen's HTML-to-video renderer: author a composition in HTML, CSS and JS, preview it in a browser, then render the same input to a deterministic MP4 locally or on AWS Lambda.
Toonflow★ 15.8kInfinite-canvas workbench for AI short-drama production, taking a novel through chapter events, script, storyboards and video nodes.
Palmier Pro★ 14.4kA Swift-native macOS video editor with generative models in the timeline and a built-in MCP server, so Claude, Codex or Cursor can edit the open project alongside you.
Hypit★ 11.8kA markup language and compiler that lets a coding agent rebuild a reference video as an editable workflow — footage, captions, B-roll and effects anchored to words rather than timestamps — and render variants in one command.
LTX-2★ 9.5kLightricks' official inference package for its LTX-2 audio-video model: a dozen pipelines covering text/image-to-video, keyframe interpolation, audio-to-video, retakes, dubbing and HDR/EXR, plus a LoRA trainer.
html-video★ 4.6kYour coding agent writes the HTML; your laptop renders the MP4