When you use Claude Code to make a video, a skill is a folder of reusable instructions and resources that helps Claude with a specific task. A video skill can guide an agent as it writes a script, draws scenes, adds narration, and renders the finished video. Public skill collections include creative workflows. Video production is one part of the wider range of available skills.
For narrated explainers and product demos, I recommend explainroo. This free, open-source kit uses the MIT license and works with Claude Code and other coding agents. Give the agent a topic or product, and explainroo helps it make and check an MP4 on your computer. The sections below explain how to use it and what to review before publishing.
What a video skill does
A video skill gives Claude instructions it can reuse for a production task. The instructions can explain how to structure a script, arrange scenes, or run rendering commands. Claude loads the relevant guidance when needed. Curated Claude skills resources show how people use skills to customize coding workflows.
Instructions alone cannot create a video. Claude also needs tools to make visuals and narration and combine them into a playable file. explainroo provides a video workflow that a coding agent runs from the command line. It works with Claude Code, Codex, Pi, OpenCode, Gemini CLI, and other agents that can run shell commands. It works best with Claude Code.
Make a video with explainroo
The quickest way to start is to give Claude Code explainroo’s setup prompt. Replace the bracketed text with your topic:
Make me a short explainer video about [your topic]. Use explainroo for it: clone the repository, read its AGENTS.md and follow the steps.After setting up explainroo, the agent creates and checks the video, then gives you an MP4 file. It creates two working files: script.md for the narration and scenes.js, a JavaScript file that describes how to draw each scene. You can time a drawing to appear on a specific word in the narration, matching the visuals to what the speaker says.
To install explainroo yourself, you need a recent version of Node.js, FFmpeg, and Chrome or Chromium. FFmpeg is a command-line tool that combines media into a video file. Run these commands from a terminal:
git clone <repository URL>
cd explainroo
npm install
node bin/explainroo.js doctor --fetchThe first setup downloads the models that generate the voice and time the narration. You do not need a graphics card. explainroo saves the finished video at videos/<name>/out/video.mp4. Its documentation explains the workflow and settings.
How explainroo creates and checks scenes
explainroo uses Kokoro, an open voice model, to read the script aloud in American or British English voices; narration is available only in English and requires no account or API key. Whisper, a speech-recognition model, listens to the recording and marks when each word is spoken. This timing lets a scene appear in sync with a specific word.
Chrome runs in the background and draws each video frame on an HTML canvas, a browser surface for graphics. You can add charts, code, and screenshots to scenes. Lines can look hand-drawn, and the project includes a large icon library. explainroo also creates background music, adds sound effects, and lowers the music under the narration. FFmpeg combines the frames and audio into an MP4.
An agent cannot watch the finished video. explainroo provides other checks: still images of scenes, a contact sheet of small frames, a layout check for cut-off or overlapping text, and a speech check for mispronounced words. Before sharing the MP4, review its pacing, visual accuracy, and product details yourself.
Choose a look, format, and pace
The video.json settings control the visual style and pace. explainroo has five looks: paper, clean, chalk, blueprint, and midnight. You can make wide videos for YouTube; tall videos for YouTube Shorts, TikTok, and Instagram Reels; portrait videos for Instagram and LinkedIn posts; or square videos. Tall, portrait, and square formats highlight captions one word at a time. The tall layout also keeps text clear of areas where those apps place buttons.
The pace setting controls the speed of the voice, pauses, and animations. If the explanation feels rushed, ask Claude to slow it down or change the setting. For example, use the wide format for a YouTube video explaining a feature, then use the tall format for a short-form post with word-by-word captions.
Create a product demo
For a product demo, tell Claude which product to show and where to find its code or website. explainroo can recreate the product’s screens with matching colors, fonts, and button labels. A mouse pointer clicks through the screens and types into fields as the narration explains what is happening.
State the demo’s goal clearly. Ask Claude to show how a user completes one task. Name the relevant screens and provide the product’s code or website. Then compare the reconstructed screens and each click with the actual product. explainroo makes narrated product demonstrations. It does not edit existing camera footage, create talking avatars or live-action footage, or include a drag-and-drop editor. To change the video, tell the agent what to revise.
Costs and practical limits
explainroo’s core features are free per video. Optional AI illustrations through OpenRouter cost per image. The coding agent is a separate service with its own pricing. If your video does not need those illustrations, the included narration and production workflow has no per-video charge.
Claude can make a draft and run explainroo’s checks, but you are responsible for reviewing the video. Check that the narration is accurate, the visuals support the explanation, and the text is readable on the target platform. The project repository contains the code and setup instructions if you want to see how the kit works.
