Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 18 additions & 0 deletions docs/custom.css
Original file line number Diff line number Diff line change
Expand Up @@ -1670,3 +1670,21 @@ pre.shiki {
.code-block-fade-overlay {
background-image: linear-gradient(to right, transparent, var(--hf-code-bg) 24px) !important;
}

/* 30 Days: a numbered syllabus, one row per lesson, the film previews on hover */
.hf-days { list-style: none; margin: 1rem 0 2rem; padding: 0; border-top: 1px solid var(--hf-ui-border); }
.hf-day { display: grid; grid-template-columns: 3.25rem minmax(0, 1fr) 10.5rem; gap: 1.25rem; align-items: center;
padding: 0.9rem 0; border-bottom: 1px solid var(--hf-ui-border); margin: 0; }
.hf-day-n { font-size: 1.75rem; font-weight: 600; line-height: 1; letter-spacing: -0.02em; color: var(--hf-text-tertiary);
font-variant-numeric: tabular-nums; }
.hf-day-title { font-weight: 600; font-size: 1rem; color: inherit; text-decoration: none; }
.hf-day-title:hover { text-decoration: underline; text-underline-offset: 3px; }
.hf-day-text p { margin: 0.2rem 0 0; font-size: 0.875rem; line-height: 1.45; color: var(--hf-text-secondary); }
.hf-day-film { width: 100%; aspect-ratio: 16 / 9; object-fit: cover; border-radius: 0.5rem; background: var(--hf-surface2); }
@media (max-width: 640px) {
.hf-day { grid-template-columns: 2.25rem minmax(0, 1fr) 6.5rem; gap: 0.75rem; }
.hf-day-n { font-size: 1.25rem; }
}

.hf-day-play { display: block; width: 100%; padding: 0; border: 0; background: none; cursor: pointer; border-radius: 0.5rem; }
.hf-day-play:focus-visible { outline: 2px solid var(--hf-brand); outline-offset: 2px; }
4 changes: 3 additions & 1 deletion docs/docs.json
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,8 @@
"pages": [
"introduction",
"quickstart",
"examples"
"examples",
"faq"
]
},
{
Expand All @@ -82,6 +83,7 @@
"guides/with-codex",
"guides/with-cursor",
"guides/mcp",
"guides/free-agents",
"guides/skills",
"guides/plugins",
"guides/design-tools"
Expand Down
157 changes: 157 additions & 0 deletions docs/faq.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,157 @@
---
title: "HyperFrames FAQ"
sidebarTitle: "FAQ"
description: "Short answers to the questions people ask most about HyperFrames: cost, which agent to use, Windows, Studio, Remotion, rendering, audio, and fonts."
---

Short answers to the questions people ask the docs assistant most often. Each
answer links to the page with the details.

## Is HyperFrames free?

Yes. The framework, CLI, and skills are open source under Apache 2.0, and
rendering on your own computer is free and needs no account. Your AI agent's
plan and optional HeyGen-hosted features, such as HeyGen voices, avatars, and
cloud rendering, bill through their own accounts. See
[what's free and what HeyGen adds](/guides/voice-and-audio#free-or-heygen-quality), and
[Make videos with a free agent](/guides/free-agents) for an agent that costs nothing.

## Can I use HyperFrames commercially?

Yes. The [Apache 2.0 license](https://github.com/heygen-com/hyperframes/blob/main/LICENSE)
lets you use, change, and distribute HyperFrames, including in commercial work,
under its terms. Media you add, such as fonts, music, or footage, keeps its own
license.

## Which AI agent should I use?

Any coding agent that supports skills works, including Claude Code, Codex,
Cursor, Gemini CLI, GitHub Copilot, Google Antigravity, IBM Bob, and OpenCode.
If you have none yet, start with Claude Code or Codex. Setup guides:
[Claude Code](/guides/with-claude-code), [Codex](/guides/with-codex),
[Cursor](/guides/with-cursor), and [other agents](/guides/skills).

## Can I use it in the Claude or ChatGPT chat app?

Yes, through the hosted HyperFrames connector. It creates and renders videos in
your HeyGen account, with nothing installed on your computer. You give up the
local project folder and Studio. Your chat product's plan decides whether you
can add a custom connector. See [Make videos in Claude.ai or ChatGPT](/guides/mcp).

## How do I install it?

For Claude Code, install the plugin:

```bash
claude plugin marketplace add heygen-com/hyperframes
claude plugin install hyperframes@hyperframes
```

For most other agents, run `npx skills add heygen-com/hyperframes` and choose
**Core Skills**. Then start a new agent session. See
[Make your first video](/quickstart).

## Do I need a HeyGen account or API key?

Not to make and render videos on your computer. You need one for HeyGen voices
and music, avatars, cloud rendering, the hosted connector, and the Studio app.
Sign in with `npx hyperframes auth login`. See
[Authentication and API keys](/guides/authentication).

## Does HyperFrames work on Windows?

The CLI and the agent skills run on Windows, including rendering from
PowerShell. The HyperFrames Studio app is available for Mac and Linux only. On
Windows, edit in the browser editor with `npx hyperframes preview`. For WSL
rendering tips, see [The render is slow under WSL2](/help#the-render-is-slow-under-wsl2).

## What is the difference between the Studio app and the browser editor?

The **Studio app** is a desktop app for Mac and Linux: a chat with your agent, a
preview, and an editor in one window, with no terminal needed. The **browser
editor** is the editor that opens from any project when you run
`npx hyperframes preview`. Both edit the same project files. See
[Get the Studio app](/studio-app) and
[Edit in the browser editor](/studio).

## How is HyperFrames different from Remotion?

Remotion videos are React components. HyperFrames videos are HTML, CSS, and
JavaScript, timed with data attributes and a paused animation timeline.
HyperFrames does not use or run Remotion. See
[HyperFrames vs Remotion](/guides/hyperframes-vs-remotion) and
[Move from Remotion to HyperFrames](/prompting/remotion-migration).

## How do I preview and render a video?

Inside the project folder, open the live preview:

```bash
npx hyperframes preview
```

Then render the MP4:

```bash
npx hyperframes render --output final.mp4
```

Rendering needs FFmpeg. Run `npx hyperframes doctor` to check your machine. See
[Render from the command line](/guides/rendering).

## How long does a render take?

It depends on length, resolution, and effects. In one measured example, a
300-frame 1920×1080 composition rendered in 9.8 seconds, and in 25 seconds with
one expensive CSS declaration left in. Use `--quality draft` for a faster review
render. See [Fix a slow preview or render](/guides/performance).

## Can it add voiceover, music, and sound effects?

Yes. Ask the agent for them in your request. With a HeyGen account it uses HeyGen
voices and a 10,000-track music catalog. Without one, voice uses Kokoro on your
computer, music is tracks you add (or local generation if you install it), and
sound effects come from a bundled set.
See [Add voiceover, music, sound effects and captions](/guides/voice-and-audio).

## Can it add captions to my video?

Yes. Give the agent your talking-head footage and ask for captions. It
transcribes the speech, times each word, and adds captions without changing the
footage. See [Add captions to a video with AI](/guides/captions-and-recuts).

## Can I use my own fonts?

Yes. Put the font file in the project and declare it with `@font-face`. A font
the project does not declare can fall back to a default face in the render, and
`npx hyperframes lint` reports it. See the
[font lint rules](/reference/lint-rules#fonts).

## Can I render in the cloud or on my own servers?

Yes. `hyperframes cloud render` renders on HeyGen's cloud and bills credits to
your HeyGen account. You can also render on your own AWS Lambda or Google Cloud
Run, or in Docker. See [Choose a rendering path](/deploy/overview).

## Does HyperFrames send my video to HeyGen?

Not when you preview and render locally: the project stays on your computer.
Cloud rendering uploads the project to HeyGen, and the hosted connector builds it
in your HeyGen account. The CLI sends usage data, such as command names and
render performance, which stays anonymous until you sign in. It never sends
project names or video content. Turn it off with `npx hyperframes telemetry disable`. Your
AI agent sends your prompts to its own provider. See
[the telemetry reference](/packages/cli#telemetry).

## Something failed. What do I check first?

Run these in the project folder and give the agent the full output:

```bash
npx hyperframes doctor
npx hyperframes lint
npx hyperframes check
```

See [Troubleshooting](/help) for specific errors, and
[Share feedback](/help#share-feedback) to report a bug.
54 changes: 54 additions & 0 deletions docs/guides/free-agents.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
---
title: "Make videos with a free agent"
sidebarTitle: "Free agents"
description: "HyperFrames is free; the agent is the only cost. Use OpenCode's free models in the Studio app, or a model on your own computer, and see how open models score."
---

import { DocsVideo } from "/snippets/docs-video.jsx";

HyperFrames costs nothing. What you pay for is the AI agent that writes the video, and that can be free too.

## Same prompt, paid model and open model

In HeyGen's [Code2Video benchmark](https://www.heygen.com/research/introducing-code2video-benchmark), models wrote
videos with the HyperFrames skills from 168 briefs and were ranked against each other. The open-weight models
(Qwen 3.8 Max, DeepSeek V4.1 Flash, Kimi K3, GLM 5.3) all landed within 50 Elo points of first place. The same prompt,
given to a paid model and to an open one:

<div className="hf-versus">
<div>
<p><strong>Claude Fable 5</strong></p>
<DocsVideo title="Product intro by Claude Fable 5" src="https://dynamic.heygen.ai/www/video/research/code2video/claude-fable-5__Product_Intro_31_ijvIoCty8.mp4" poster="https://dynamic.heygen.ai/www/video/research/code2video/claude-fable-5__Product_Intro_31_ijvIoCty8.mp4/ik-thumbnail.jpg?tr=so-3" />
</div>
<div>
<p><strong>DeepSeek V4 Flash</strong></p>
<DocsVideo title="Product intro by DeepSeek V4 Flash" src="https://dynamic.heygen.ai/www/video/research/code2video/deepseek-v4-flash__Product_Intro_31_wJgiSCn0jw.mp4" poster="https://dynamic.heygen.ai/www/video/research/code2video/deepseek-v4-flash__Product_Intro_31_wJgiSCn0jw.mp4/ik-thumbnail.jpg?tr=so-3" />
</div>
</div>

The top models are still better at the hard parts: timing, composition, finish. For a first draft, a social post or
learning the tool, an open model is often enough.

## In the Studio app: OpenCode

[Connect an agent](/studio-app/connect-your-agent) and choose **OpenCode**. The app installs it, and it runs free
models with no sign-in. It needs **Full access**, and the app asks to turn it on the first time.

## On your own computer

OpenCode can also use a model running locally, so nothing leaves your machine and nothing is billed. With
[Ollama](https://ollama.com) running, OpenCode finds its models by itself; pick one that supports tools. See [OpenCode's model docs](https://opencode.ai/v2/docs/models).

Local models are slower and weaker than the hosted ones, so start with short videos.

## What else is free

Rendering, previews, Kokoro voices, background removal and the bundled sound effects run on your computer at no
cost, and so does transcription when your own agent does it. The Studio app is free with a free HeyGen account. HeyGen's voices, avatars and generated clips use HeyGen credits;
everything else in a video is code. See [free or HeyGen quality](/guides/voice-and-audio#free-or-heygen-quality).

## Related topics

- [Connect your agent](/studio-app/connect-your-agent)
- [Use HyperFrames with Claude Code](/guides/with-claude-code)
- [The Code2Video benchmark on Kaggle](https://www.kaggle.com/benchmarks/heygen/code2video)
Loading
Loading