Guide · Agent skill
How to use the Openscreen screen recording CLI agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
This guide explains how to install and use the cli-anything-openscreen skill, which lets an AI agent edit existing screen recordings through a command-line tool. It is written for readers who have never used agent skills. By the end you will know what the skill does, how to add it to your agent, what it needs, and how to ask for edits in plain language.
Quick start
- 1Download and unzip the skill. Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or your project's .claude/skills/ folder. In the Claude apps, you can upload it under Settings → Capabilities → Skills instead.
- 2Install Python 3.10+ and ffmpeg. Confirm you have Python 3.10 or newer and that ffmpeg is installed on your system. The tool will not render video without ffmpeg.
- 3Install the command-line package. Run pip install cli-anything-openscreen. Then run cli-anything-openscreen --help to confirm the command works.
- 4Create a project from a recording. Ask the agent to create a project from your video, or run project new with -v for the video and -o for the project file name. Projects are saved as .openscreen JSON files.
- 5Request edits in plain language. Describe the zoom, speed, trim, crop, annotation or background changes you want. The agent uses the matching command groups and the undo history if you change your mind.
- 6Preview and export. Ask for a quick preview to check the edits, then ask for an export. Verify the output file exists and probe it before sharing.
What an agent skill is and what this one does
An agent skill is a folder with a SKILL.md file that tells an AI agent how to perform a specific task with a specific tool. This skill covers cli-anything-openscreen, a command-line interface for Openscreen, a screen recording editor.
Once installed, the agent can turn raw screen captures into polished demo videos. The source lists zoom effects, speeding up idle sections, cutting out parts, cropping the frame, adding text annotations, setting backgrounds and aspect ratios, and rendering the final video with ffmpeg. The tool also keeps an undo/redo history and can produce preview bundles of frames and a short clip so the agent can check its own edits.
Install the skill in Claude
Download the ZIP of the skill from this page and unzip it. For a personal setup, place the folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
In the Claude apps you can instead upload the skill under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Project: .claude/skills/ inside your project folder
- Claude apps: Settings → Capabilities → Skills
Install the command-line tool and its requirements
The skill drives a command-line program, so that program must be installed too. The source lists these prerequisites: Python 3.10 or newer, and ffmpeg installed on your system. Install the tool with pip using the command below.
Verify the install by running the help command. Running the tool with no arguments opens an interactive mode where you type help for commands and quit to exit.
- Python 3.10+
- ffmpeg installed on your system
- pip package: cli-anything-openscreen
pip install cli-anything-openscreenAsk the agent for edits in plain language
You do not need to type the commands yourself. Describe the result you want, such as trimming the start of a recording, adding a zoom on a button click, or exporting a 1:1 version. When the request matches the skill, the agent runs the cli-anything-openscreen commands for you.
The agent works from times in milliseconds and coordinates normalized from 0 to 1, so give timings in seconds if you like and let the agent convert them. Be specific about which part of the recording you mean.
- Trim the first 4 seconds and the last 2 seconds of demo.mp4.
- Zoom to 2.2x on the top-left area between 3 and 6 seconds.
- Speed up the idle section from 10 to 25 seconds to 2x.
- Add the text "Click Save" near the bottom middle from 8 to 11 seconds.
- Make a 9:16 export with a dark gradient background and export it as mp4.
Commands the agent can run
The tool is organized into groups. Each group has list and remove commands for its regions, and the project, media, export, preview and session groups handle files and state.
Zoom depths map to magnification: 1=1.25x, 2=1.5x, 3=1.8x, 4=2.2x, 5=3.5x, 6=5.0x. Valid speeds are 0.25, 0.5, 0.75, 1.25, 1.5, 1.75 and 2.0. Aspect ratios are 16:9, 9:16, 1:1, 4:3 and 4:5.
- project: new, open, save, info, set-video, set
- zoom: list, add, remove
- speed: list, add, remove
- trim: list, add, remove
- crop: get, set
- annotation: list, add-text, remove
- media: probe, check, thumbnail
- export: presets, render
- preview: recipes, capture, latest
- session: status, undo, redo, save, list
cli-anything-openscreen project new -v recording.mp4 -o project.openscreen
cli-anything-openscreen --json project infoCheck your own work with previews and exports
The preview group renders a low-res bundle that the agent can inspect. A quick bundle typically contains a low-res clip, sampled frames, a hero frame and a summary.json with trim, zoom and speed facts. Previews are static, not live, so they are a review aid rather than a live editor.
After rendering, the source advises verifying that the output exists and probing it. The source also notes that the tool writes project and output files to disk, so keep copies of original recordings.
- Run preview capture with the quick recipe to check edits.
- After export render, confirm the file exists and run media probe on it.
- Keep backups of original recordings before editing.
cli-anything-openscreen preview capture --recipe quick
cli-anything-openscreen export render output.mp4What people use it for
- Turn a raw screen capture of a product walkthrough into a short feature demo with zoom on key clicks.
- Remove dead time from a tutorial recording by speeding up idle sections and trimming the ends.
- Produce the same demo in several aspect ratios, such as 16:9 for a website and 9:16 for a vertical post.
- Add text annotations at set times to explain each step of a recording.
- Run scripted, repeatable edits on a batch of recordings with the same settings.
Frequently asked questions
How do I install cli-anything-openscreen?
Install it with pip using pip install cli-anything-openscreen. You also need Python 3.10 or newer and ffmpeg installed on your system. Then download the skill ZIP and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder.
What do I need to run the Openscreen CLI skill?
The source lists Python 3.10+, ffmpeg installed on the system, and the pip package cli-anything-openscreen. The skill runs scripts, so it is best used on a machine where you are comfortable with that.
Is cli-anything-openscreen free to use?
The project is listed under the Apache-2.0 licence. Check the licence file in the source repository for the full terms before commercial use.
Can this skill record my screen?
No. The skill edits existing video files and does not capture screens, because no capture commands are listed.
Why does the agent fail or report an error?
Check that ffmpeg is installed and that Python is 3.10 or newer. The source advises checking return codes, where 0 means success and non-zero means error, and reading stderr for messages in non-JSON mode.
Does it need an internet connection?
The listing says the skill does not use the network. It works on local video files and writes project and output files to disk.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you want an agent or a scripted workflow to turn existing screen recordings into demos with zoom, trims, speed changes, crops and text, and you are comfortable with a command-line tool. Pick a graphical video editor instead if you need live editing or a visual timeline, since this workflow is command-line and static preview based.
Download the skill