Studio · engine · what is which
HyperFrames AI
Two different things share the name HyperFrames. One is a hosted AI workspace that turns a prompt or a photo into a video or an image in the browser. The other is an open-source engine you run yourself. This page explains both, so you can pick the one that matches what you actually searched for.
$4.99 one-time · 50 credits · no subscription · Start with the free browser tools or one 50-credit pack. Pay for generation only when you need new footage.
- Credit cost is shown before you submit
- Commercial-use license on paid plans
- Free browser tools for format changes
Presets
Start from a tested preset
Click “Use this preset” to put the prompt into the Studio on this page, or copy it and edit the details yourself. Each example is an original concept, not a licensed character.
First Clip
A single stable shot that is the cheapest way to test the model.
View prompt
Create a four-second cinematic clip: a lone figure walking through fog at dawn along an empty road, cold blue light breaking warm at the horizon, slow dolly forward, realistic mist. One continuous shot, stable geometry, natural motion, no text, no logos, no watermarks.
Product Clip
A single product on a plinth with one slow move and clean light.
View prompt
Create a five-second product clip: a matte ceramic perfume bottle on a dark stone plinth, lit by one warm beam from the side, slow orbit around the bottle, faint dust in the light, deep shadows. Studio product photography, one continuous shot, no text, no logos, no watermarks.
Try it
Open the Studio with this prompt
The prompt is already in the box. Replace the example details with yours, upload a photo if the workflow needs one, and check the credit cost before you submit.
- 1Choose a prompt or a preset above.
- 2Upload an authorized photo if the workflow asks for one.
- 3Generate one short draft, review it, then refine one variable.
Starter prompt
Create a four-second cinematic clip: a lone figure walking through fog at dawn along an empty road, cold blue light breaking warm at the horizon, slow dolly forward, realistic mist. One continuous shot, stable geometry, natural motion, no text, no logos, no watermarks.
HyperFrames AI
Generation Controls
These are the shared controls across all three workflow lanes.
Uses the faster route when available.
Include synchronized audio when the provider supports it.
Task Notifications
Useful for longer queue times. We keep the current task switch separate from your future default.
We send the prompt, task page link, and a direct download shortcut when the render finishes.
Save this as your default so new tasks start with notifications already enabled.
Nothing generated yet
Start with one real test.
$4.99 one-time · 50 credits · no subscription. Start with the free browser tools or one 50-credit pack. Pay for generation only when you need new footage.
Guide
How to get a usable result
The short version: keep the change small, describe what must stay the same, and review the first draft before spending more credits.
What HyperFrames AI usually means
When people type “HyperFrames AI” they usually want one of two things: a browser tool that generates video or images from a prompt, or the open-source HyperFrames framework that renders HTML into video from a command line. The names collide, the search results collide, and it is easy to land on the wrong one.
This site, hyperframes.net, is the first thing. It is an independently operated hosted workspace with five generation workflows: text to video, image to video, reference to video, text to image and image to image. It does not require an install and it runs in the browser.
The second thing is the HyperFrames open-source framework published by HeyGen, released under the Apache 2.0 licence and driven from the command line. It composes HTML, CSS and JavaScript frames and renders them into an MP4 deterministically. It is a developer tool, not a hosted product, and this site is not its homepage.
What the hosted HyperFrames AI studio does
The Studio is a generation workspace. You describe a shot, or upload a photo, pick a duration and resolution, and the workspace renders a clip you can preview and download. The five workflows cover the common cases: a scene with no source image, a still that should move, a reference clip whose style and motion you want to keep, a single image, and an edit of an existing image.
It is deliberately not a timeline editor. There is no multi-track cutting, no keyframes, no audio mixing and no colour suite. The value is in creating a shot that did not exist, quickly and without a shoot. Cut, caption and score the result in whatever editor you already use.
Everything is server-side, so a phone or a low-power laptop can drive it. There is nothing to install, no model to download, and no GPU to babysit. That is the practical difference between the hosted workspace and the open-source engine.
HyperFrames AI versus the open-source engine
The engine is deterministic and free to self-host, and it is the better answer if you need programmatic rendering inside a pipeline: a build step that turns a template plus data into hundreds of MP4s, a reproducible render in CI, or HTML that you already generate. You supply the compute and you control the environment.
The hosted Studio is the better answer if you want a shot, not a build system. You trade control over the render environment for not having to run one, and you pay per generation instead of per GPU hour. Most people searching “HyperFrames AI” want this second thing.
A useful rule: if your question starts with “how do I install”, “how do I call it from a script” or “can I run it offline”, you want the open-source framework. If it starts with “can I just make this clip now”, you want the Studio.
- Runs in the browser: hosted Studio. Runs from a terminal: open-source engine.
- Pay per generation in credits: hosted Studio. Pay in compute and setup time: self-hosted engine.
- Deterministic, scriptable HTML to MP4: open-source engine. Prompt or photo to footage: hosted Studio.
- No install and no GPU: hosted Studio. Full control of the render environment: open-source engine.
What a clip costs
Generation runs on credits. A four-second 720p Fast clip starts at 40 credits, and the smallest one-time pack is 50 credits for $4.99, which covers that first test with a few credits left over. Larger packs lower the effective price per credit, and the Studio shows the exact cost before you submit a job.
The five browser tools are a separate, free path: video converter, video compressor, video thumbnail extractor, video to GIF and image compressor. They run locally in the browser, do not consume credits, and cover the finishing steps that never needed a generation model in the first place.
The practical way to spend less is to iterate on a short, cheap clip until the prompt is right, then render the longer or higher-resolution version once. Regenerating a finished-looking clip at full length to fix a bad prompt is the most common way to waste credits.
Writing a prompt that survives the first render
One subject, one setting, one camera move. A prompt that asks for three locations and two characters in five seconds will smear, because the model has to invent a cut it was never asked to make. “A slow drone push over a rocky coastline at sunset, warm low light, soft haze” is a shot. “A travel montage across Europe” is not.
Add stability language: one continuous shot, stable geometry, natural motion, consistent lighting. Name what you do not want as well, since text in frame and watermarks are the two artefacts that ruin an otherwise usable clip. Then reuse the same palette and camera language across shots so the results cut together.
For image to video and reference to video, the source image does most of the work. Start from a sharp, well-lit frame with a clear subject, and describe the motion rather than the content: “slow push in”, “gentle parallax”, “fabric moving in a light breeze”.
Who should use which
Solo creators, marketers and small teams should start with the hosted Studio. The bottleneck for them is usually the first usable asset, not render throughput, and paying a few dollars for a clip is cheaper than a morning of setup.
Developers with an existing render pipeline, or anyone producing the same template at volume, should start with the open-source engine. It removes per-render cost and makes the output reproducible, which matters when a video is part of a build.
Plenty of teams end up using both: the Studio for the exploratory shots nobody wants to script, the engine for the repeatable ones. That combination is the honest answer to “HyperFrames AI or HyperFrames framework” — they solve different halves of the same job.
FAQ
Questions people ask before they generate
Is HyperFrames AI the same as the open-source HyperFrames framework?
No. This site is a hosted generation workspace: you write a prompt or upload a photo and get a video or image in the browser. The open-source HyperFrames framework is an Apache 2.0 project driven from the command line that renders HTML into MP4. They share a name, not a codebase.
Is hyperframes.net operated by the authors of the open-source framework?
No. hyperframes.net is an independently operated hosted workspace. The open-source framework lives in its own GitHub repository and is maintained by its own authors. If you need the framework itself, go to that repository rather than this site.
Do I need to install anything to use HyperFrames AI?
No. The hosted Studio runs in the browser and renders server-side, so there is nothing to install and no GPU to configure. The open-source framework is the opposite: it is installed locally and rendered on your own machine.
How much does a generated clip cost?
A four-second 720p Fast clip starts at 40 credits, and the smallest one-time pack is 50 credits for $4.99. Longer and higher-resolution renders cost more, and the Studio shows the exact cost before you submit the job.
Can I use the free browser tools without an account?
Yes. The video converter, video compressor, video thumbnail extractor, video to GIF and image compressor run locally in the browser and do not consume generation credits. They cover format changes, compression, posters and GIFs.
Which one should I pick if I already have a render pipeline?
Use the open-source framework when the render has to be reproducible inside a build or scripted at volume. Use the hosted Studio for exploratory shots and one-offs, where setting up a render environment would cost more than the generation.