Video Editing and Creation with Claude · The HyperFrames and Video Use Guide.

Most people think about video editing like this: you shoot, open software like Premiere or CapCut, sit for hours cutting mistakes, removing silences, adding subtitles, and finally adding effects. Every step requires skill, and some steps require paid software. In session 24 of the AI Master Club, we turned this entire process into a conversation. Instead of working against a mouse and screen, you simply talk to Claude (or Codex) and tell it what to do. It writes the code behind the scenes and performs all the editing. Adding subtitles, animations, cutting silences, integrating a logo, animating images, everything by command. The two tools that work the magic are HyperFrames and Video Use, and both are free. HyperFrames is a tool by HeyGen (the company that makes avatars) that adds subtitles, animations, and visual effects on top of videos. Video Use is a new plugin that enables the editing work itself: cutting, synchronization, removing silences. Connect both to Claude as skills, and you get a complete editing studio that works by speech. This guide is the entry point: what each tool does, how to install them in a single line, and how to work correctly to get a polished product without touching editing software.

What you will have by the end of the guide

• What is the difference between HyperFrames and Video Use, and why you need both
• How to install both skills for Claude or Codex in a single line from GitHub
• Adding Hebrew subtitles, animations, and visual effects to an existing video
• Animating a static image, carousel, or presentation, without a graphic designer
• Building a promo video from a website link only
• Automatic cutting of silences and mistakes from the video
• When to use HeyGen for avatar creation, and what is the difference from Higgsfield
• How to manage Claude correctly during the editing process, and how to package everything as a skill
• All the tools, links, and questions that came up in the session

The shift from a chat that answers to a chat that executes

HyperFrames · Subtitles, animations, and effects

Video Use · The video editor itself

Localhost: Often when Claude creates a preview, it places it on Localhost (for example localhost:3000). This is like a website that exists only on your computer, available only to you and to no one else. This is how you test the product before downloading it as a final MP4 file.

How to install · Two skills in one line

What you need to work: A paid subscription to Claude Code or Codex (minimum 20 dollars a month), because they are the ones that run the tools and write the code. In addition, if you want to add subtitles to the video, you need one API Key from ElevenLabs (possible on a free account). ElevenLabs is a voice generation tool, but here it is used to transcribe the video and understand what was said in it, in order to know where to place the subtitles.

Prompt for animating an image or graph

I am attaching a static image [graph / slide / infographic]. I want you to use HyperFrames and animate it. Create an animation that shows [for example: how the percentage rises gradually and how the graph is built]. Keep the original design and text, including Hebrew. At the end, give me a preview on Localhost and then a downloadable file.

Prompt for full video editing

I opened a folder named [folder name] with my videos. I want you to edit the video: remove all silences and mistakes (ums), add Hebrew subtitles synchronized to what I am saying, and integrate the logo [location] without covering my face. Use Video Use for cuts and HyperFrames for subtitles and visuals. Show me a work plan first, and at the end give me an MP4 file to download.

Managing Claude like a new employee

The Avatar · HeyGen behind the scenes

The AI Master Club opens the door for you: a live session every week, recordings, and a community. 47 NIS per month, cancel anytime.

HyperFrames is not a Hollywood video generation tool. It does not take an image and turn it into a cinematic video. It dresses subtitles, animations, and visual effects on top of existing videos, or animates 2D images graphically. For video generation from scratch (animating a driving car, a realistic scene), other tools are needed: Veo and Omni through Gemini and Flow, or Higgsfield. These are tools that cost money and use different technology. HyperFrames and Video Use are free, and they are for editing and visuals.

Claude or Codex · Do you need both?

From Higgsfield into the video

All tools from the session

The skill that adds subtitles, animations, and visual effects to videos. Install via the Code button, copy and paste to Claude. Open source, no payment for rendering

The official website with the catalog: all types of animations, effects, transitions, and product demos that can be created

The skill that performs the editing itself: cutting, removing silences and mistakes, synchronization, and burning subtitles. Install exactly like HyperFrames

The platform where HyperFrames lives, and where you create an avatar and digital twin from three minutes of video

Requires one API Key (free option available) so the system can transcribe the video and know where to place the subtitles

The agent that writes the code and manages all the editing. Install the skills in it and work by speech

The alternative to Claude Code. Anyone with a ChatGPT subscription already gets it. Works with the exact same skills

A tool for generating realistic video from scratch (paid). Can be connected to Claude and integrate the outputs inside the edited video

Google's studio where Omni lives, for creating video and avatars. Learned in session 23 (link to guide below)

The basic subscription gives credits for creating video with Omni. The avatar created in Flow is also available here

A source for inspiration: images, graphs, slides, and fonts that can be fed into HyperFrames for it to animate them

Questions that came up in the session

HyperFrames is a tool inside HeyGen, but everything it does is done in code. When you connect it to Claude as a skill, Claude is the one that writes the code and runs the tool. This is how you can work with HyperFrames for free from within Claude, animate images and videos, and add subtitles and animations, without working directly inside the platform.

You do not need to know how to code at all. Claude writes the code by itself. It does help to know basic editing terms (zoom in, transition, subtitle) to guide it precisely. The new skill of today is knowing how to direct: say what you want, and do not cut manually.

As long as you are in the same chat, the same folder, and the same process, it remembers everything it did. And if it is a recurring process, you can package it as a skill. Next time it will pull the skill automatically and execute everything without you explaining again.

Yes, Hebrew subtitles work great. You need to connect an ElevenLabs API Key so the system can transcribe the video and know what was said in it. Feeding written text to an avatar in HeyGen is still not precise in Hebrew, but the subtitles themselves in HyperFrames are excellent.

Yes. You can feed a static image, graph, slide, or a set of images for a carousel, and HyperFrames animates them. You can also give it a website link and it builds a promo video from it. A filmed video is not required at all.

Higgsfield creates video from scratch, including realistic videos, and costs money. HyperFrames does not create realistic video, but dresses subtitles, animations, and effects on existing videos, or animates 2D images, and it is free. You can connect Higgsfield to Claude and integrate its outputs inside the video being edited in HyperFrames.

If you do not reach the quota of one of them, there is no point in both. Both work with HyperFrames and Video Use because these are skills. If you already have a ChatGPT subscription, you have Codex. If you are used to Claude, work with it.

And what is in the next session

The full recording of session 24, along with all links and explanations, is available to club members in the content library. The only thing limiting the work here is your imagination. Try HyperFrames and Video Use, play around, and you will discover that it is free, simple, and works. Share your outputs in the community, it helps everyone advance.

Frequently Asked Questions

How does Claude Code actually connect here? I thought everything is done in HeyGen.

HyperFrames is a tool inside HeyGen, but everything it does is executed in code. When connected to Claude as a skill, Claude is the one that writes the code and runs the tool. This way, you can work with HyperFrames for free from within Claude, animate images and videos, and add subtitles and animations, without working directly inside the platform.

Do I need to know how to program or edit video?

You do not need to know how to program at all. Claude writes the code by itself. It does help to know basic editing terms (zoom in, transition, subtitle) to direct it precisely. The new skill today is knowing how to direct, which means saying what you want rather than cutting manually.

Does it remember the process, or do I need to explain it again for every video?

As long as you are in the same chat, the same folder, and the same process, it remembers everything it did. If this is a recurring process, you can package it as a skill. Next time, it will retrieve the skill automatically and execute everything without you having to explain again.

Do Hebrew subtitles work?

Yes, Hebrew subtitles work excellently. You need to connect an ElevenLabs API Key so the system can transcribe the video and know what was said. Feeding written text to the avatar in HeyGen is still not accurate in Hebrew, but the subtitles themselves in HyperFrames are excellent.

Can I use just an image, without filming a video?

Yes. You can input a static image, a graph, a slide, or a set of images for a carousel, and HyperFrames animates them. You can also give it a website link and it builds a promo video from it. Filmed video is not required at all.

What is the difference between HyperFrames and Higgsfield?

Higgsfield creates video from scratch, including realistic videos, and costs money. HyperFrames does not create realistic video, but rather overlays subtitles, animations, and effects on existing videos, or animates 2D images, and it is free. You can connect Higgsfield to Claude and combine its outputs inside the video you edit in HyperFrames.

Should I keep both Claude and Codex?

If you do not reach the quota of either of them, there is no point in having both. Both work with HyperFrames and Video Use because these are skills. If you already have a ChatGPT subscription, you have Codex. If you are used to Claude, work with it.