Claude can edit your videos

Edit Your Raw Footage Into a Ready-to-Upload Video With One Prompt

#The video-use skill for Claude Code: setup guide + master prompt

Hey, thanks for commenting "Skill".

Here's everything you need to go from a folder of raw clips to a finished, upload-ready video using video-use, an open-source video editing skill for Claude Code.

#What Is video-use?

video-use is a free, open-source skill that lets Claude Code edit your videos end to end. You put your raw footage in a folder, give Claude a prompt, and get a finished final.mp4 back.

It works for talking heads, tutorials, montages, travel videos, interviews and more.

What it handles for you:

• Cuts filler words (umm, uh, false starts) and dead space between takes
• Removes awkward silences and bad clips
• Color grades every segment
• Adds smooth audio fades at every cut, so there are no pops
• Burns in subtitles (customizable style)
• Adds your assets: logos, B-roll, music, overlays
• Checks its own output at every cut before showing you the result

How it works under the hood: Claude never actually "watches" your video. The skill transcribes the audio and turns the footage into a lightweight text-and-image map, so Claude can make editing decisions like a real editor, then uses ffmpeg to render the final cut.

GitHub repo: https://github.com/browser-use/video-use

#What You'll Need

• Claude Code installed on your computer
• ffmpeg: Claude will help install it during setup
• An ElevenLabs API key, used for transcription. Get one at elevenlabs.io (Settings → API Keys)
• Your raw footage, plus any assets you want in the video

#Step 1: Set It Up (One Time Only)

You don't need to know any code. Just let Claude do it.

1. Open your terminal and start Claude Code:

   claude

2. Paste this prompt:

   Set up https://github.com/browser-use/video-use for me.
   Clone the repo, follow its install.md to install dependencies and ffmpeg,
   register it as a Claude Code skill, and set up my ElevenLabs API key
   (ask me to paste it when you need it). Then read SKILL.md and the
   helpers/ folder so you understand how to use it.

3. Paste your ElevenLabs API key when Claude asks for it.

That's it. The skill is installed.

git clone https://github.com/browser-use/video-use ~/Developer/video-use
ln -sfn ~/Developer/video-use ~/.claude/skills/video-use

Then add your ELEVENLABS_API_KEY to a .env file in that folder, and make sure ffmpeg is installed.

#Step 2: Edit Your Video in 3 Steps

#1. Put everything in one folder

Create a folder and drop in:

• All your raw footage clips
• Any assets you want used: logo, intro/outro, B-roll, background music, images

Tip: Give your assets clear names like logo.png, bg-music.mp3 or broll-office.mp4. It helps Claude know what each file is for.

#2. Open Claude Code in that folder

cd path/to/your/video-folder
claude

Then call the skill:

/video-use

#3. Paste the master prompt below

#The Master Prompt

Copy this, fill in the parts in [brackets], and paste it in:

Please fix the layout, composition, and visual quality issues from the previous render. Right now, all motion graphics and overlays are cluttered in the top-right corner, and the output does not feel premium. 

Rewrite/adjust the video processing pipeline based on the following comprehensive creative and technical guidelines:

### 1. Canvas Setup & Full-Canvas Utilization
- Target Aspect Ratio: Landscape (16:9 widescreen, e.g., 1920x1080).
- Scale/crop the base video to fit 16:9 properly without stretching or pillarboxing.
- Stop crowding graphics into one corner. Distribute text, UI mockups, code snippets, and motion graphics dynamically across the entire 16:9 canvas (lower-thirds, left/right splits, center callouts) based on visual balance and storytelling context.
- Safe Area Boundaries: All floating elements must fit strictly within the 16:9 frame without bleeding off-screen.

### 2. Spatial Depth & 3D Layering (Foreground vs. Background)
- Implement proper subject-to-graphic z-index separation:
  - Behind the Subject: Place ambient background elements—such as subtle tech wireframes, grid tracks, data streams, or soft accent glows—*behind* my cutout/subject to create realistic 3D depth.
  - In Front of the Subject: Place interactive UI cards, app badges, or focal overlays in the foreground. Let them move naturally around the frame edge without obscuring my face, eyes, or expressions (unless intentionally cutting to a full-screen diagram/b-roll transition).

### 3. Face Awareness & Overlay Rules
- Place lower-thirds, captions, and side-by-side graphics in empty zones to keep my face completely visible and unobstructed during talking-head segments.
- Full-Screen Exceptions: Cover the face completely *only* during full-screen B-roll cuts, major code/architectural diagrams, or intentional full-screen graphic transitions.

### 4. Interactive Storytelling & Motion Graphics
- Synchronize motion graphics dynamically with audio cues and script topics (e.g., when mentioning store migration or push tokens, animate sleek glassmorphic store approval cards smoothly into view).
- Pacing & Transitions: Use smooth cubic-bezier easing, subtle spring physics, and high-end tech-agency aesthetics. Avoid abrupt pop-ins, cheap animations, or harsh primary colors.

### 5. Premium Caption Styling
- Aesthetic: Modern, high-end YouTube tech / agency style using sleek sans-serif typography (e.g., Inter, SF Pro Display, or Montserrat).
- Pacing: 2–4 words maximum per frame with dynamic word-by-word or phrase highlights. Avoid long walls of text.
- High-contrast text with soft drop-shadows or subtle dark pill backdrops for high legibility, highlighting key technical terms with subtle accent glows.

### 6. Research & Creative Liberty (Tech Niche Standards)
- You have full liberty to look up, evaluate, and adopt high-performing visual formats, composition styles, and editing techniques used by top-tier tech creators, design studios, and product launch reels (e.g., Apple keynotes, Marques Brownlee, Fireship, or modern SaaS demo reels).
- Mix and adapt the best compositing patterns (interactive floating cards, kinetic typography, split-screen diagrams) that best match the rhythm and narrative flow of the video.

Please update the script/pipeline to enforce these precise framing, positioning, depth-layering, and graphic quality constraints before generating the final output.

Then sit back while Claude edits your video end to end.

Your finished video will be in an edit/ folder inside your footage folder.

#Tips for Better Results

• Record clean audio. Transcription quality drives the edit, so a decent mic makes a big difference.
• Pause between takes. A clear gap before restarting a line helps the skill find the best take.
• Be specific in your prompt. Platform, length and style all change the result.
• Iterate. Not happy with something? Just tell Claude: "make the intro shorter" or "remove the clip where I'm looking away."
• Save your style. Once you like a result, reuse the same master prompt every time for a consistent look.

#Building Something Bigger Than a Video?

Tools like this show how fast you can ship when you pair the right idea with the right tech.

At Clentr, that's what we do for founders: we take messy ideas and turn them into real software products and scalable MVPs. If you've got an app or product idea you want to build, book a free call with our team and we'll help you figure out what to build first.

Follow @theshaikazad for more AI tools, tech breakdowns, and the founder journey.

Peace.