Free install guide · open-source tools

Let AI edit your videos with Video Use

Video Use is a free, open-source tool from the team behind Browser Use. You drop raw footage in a folder, tell Claude Code what you want in plain English, and get a finished video back. Here's how to set it up, step by step.

What it costs: the tool is free (MIT license). You need a coding agent: Claude Code comes with Claude Pro, about $20/month, or use Cline, a free open-source agent. Transcription uses an ElevenLabs API key. ElevenLabs has a free tier, and heavier use needs a paid plan. Check current pricing before you commit.

What you need

1Create a GitHub account

GitHub is a website where code lives, like a shared drive for code. Git is the tool that copies a project from GitHub onto your computer. If you want a point-and-click way to do that, install GitHub Desktop; it handles Git for you. Your coding agent can also install Git for you.

2Connect GitHub to your agent

In the Claude app: Settings → Connectors, search for GitHub, and connect it. That lets Claude work with GitHub directly. Other agents can use the GitHub CLI or the GitHub MCP server.

3Copy the Video Use repository

Go to github.com/browser-use/video-use. Click the green Code button and copy the URL. Make sure the owner is browser-use: popular projects attract fake copies.

4Ask your agent to install it

Open Claude Code and paste:

Install this GitHub repo for me: https://github.com/browser-use/video-use

The agent clones the project, installs what it needs (Python packages and ffmpeg), and sets Video Use up as a skill. If something fails, paste the error back and ask it to fix it.

5Add your ElevenLabs key

  1. In ElevenLabs, open API Keys and click Create Key.
  2. Give it Speech-to-Text access only. It doesn't need anything else.
  3. Copy the key and keep it somewhere safe.
Never paste your API key into the chat. When the agent asks for it, put it in the project's .env file yourself (ELEVENLABS_API_KEY=…). Anything you paste into a chat can end up in logs and transcripts.

6Edit your first video

Point the agent at your video and just ask in plain English. No special prompts:

Edit this video. Cut the dead space, add visuals and some motion graphic animations.

It transcribes the audio word by word, plans the cuts, renders, and checks its own work before showing you. Then iterate: "cut the part where I…", "pop up a graphic when they say…", "make the captions bigger". Expect a back-and-forth, not one perfect pass.

Before you run any open-source tool

CheckWhy
Official repo onlyCheck the owner name. Fakes copy popular projects.
Read the licenseMIT means you can use it commercially, so you can sell the edit, not the code.
Recent activity and open issuesShows whether it's maintained and what's broken.
Stars ≠ safetyStars show popularity, not that the code is safe.
Test on copiesBack up first. Never run it on your only copy of a file.

The Quota newsletter

Get the next tool first.

New open-source tools, builds you can copy, and the AI news that actually matters. Free. Unsubscribe anytime.