HyperFrames is an open-source framework from HeyGen that turns HTML, CSS, and seekable animations into deterministic MP4 videos. Compositions are plain HTML files with data attributes — no React, no proprietary timeline format. This makes it a natural fit for AI coding agents, which already write HTML fluently. The framework supports GSAP, CSS animations, Lottie, Three.js, Anime.js, and custom animation runtimes through an adapter system. With 29k+ GitHub stars and rapid growth since its April 2026 launch, HyperFrames is becoming the default way agents produce video content programmatically. Apache 2.0 licensed, no per-render fees.
HyperFrames solves a specific problem: AI coding agents can write HTML effortlessly, but turning that HTML into video has historically required separate tooling, proprietary formats, or manual rendering steps. HyperFrames collapses that gap. Write an HTML composition with data attributes for timing and transitions, run the CLI, get a deterministic MP4. No React dependency, no proprietary timeline schema, no cloud rendering required.
The framework is production-tested at HeyGen and has been adopted by teams at tldraw, TanStack, and others. At 29k+ GitHub stars since its mid-April 2026 launch, it’s one of the fastest-growing open-source projects in the AI video tooling space.
HTML-native compositions: Video scenes are HTML files. Use any CSS layout, any animation library, any web technology you already know. Data attributes control frame timing, transitions, and scene boundaries. This means agents don’t need to learn a video DSL — they write the same HTML they already produce.
Deterministic rendering: The same HTML input produces the same frames and output every time. This makes HyperFrames suitable for CI pipelines, automated quality checks, and regression testing of generated video content. Non-deterministic rendering is the enemy of automation.
Animation adapter system: Plug in GSAP for timeline-based animation, Lottie for vector animations, Three.js for 3D scenes, CSS animations for simple transitions, or bring your own runtime. The adapter layer normalizes seeking and frame capture across engines.
Agent skills: Pre-built skills teach coding agents the HyperFrames production loop. Compatible with Claude Code, Cursor, Gemini CLI, Codex, and any agent that supports skills or MCP.
Agent-generated video content: Let coding agents produce explainer videos, product demos, or social media clips by writing HTML compositions and rendering them locally.
Automated video pipelines: Integrate HyperFrames into CI/CD to generate video artifacts from templates — changelogs, release notes, data visualizations rendered as MP4.
Rapid prototyping: Sketch a video concept in HTML faster than any traditional video editor. Iterate in code, render in seconds.
Developers and agent builders who want programmatic video generation without cloud dependencies or per-render costs. Teams already using HeyGen who want local rendering. Anyone building AI agents that need to produce video output as part of their workflow.
Open-source game development framework powered by Claude Code — autonomous agent loop for designing, iterating, and playtesting games using natural language.
OpenAI's AI image generation model that creates and edits realistic images from natural language descriptions.
AI voice platform that generates ultra-realistic speech, voice cloning, and audio content from text with human-like expressiveness.