paper-reel skill
Claude Code skill that draws vertical 9:16 explainer reels in a cut-paper style: script, voice and MP4 from one prompt
- 138 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Runs local Node, Chromium and ffmpeg processes and downloads speech models, so it installs packages and writes files to disk
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/paper-reelDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs local Node, Chromium and ffmpeg processes and downloads speech models, so it installs packages and writes files to disk
This is third-party code. Review the repository files before installing.
What it does
You name a topic, Claude writes six or seven lines and the `reel export` command renders an MP4. Visuals are drawn in code on an HTML canvas: flat textured paper shapes, a box mascot in a new hat every scene, a headline sign, rubber stamps and counters. Captions pop in one paper scrap per word. The voice is the clock. Each spoken line becomes a scene, Kokoro voices it and whisper.cpp times every word, and props cue off those words: a stamp slams on "free", a counter runs up on "stars". Everything runs locally on a Mac with Kokoro, whisper.cpp, Playwright and ffmpeg, with no keys or accounts. Export gives 1080 by 1920 at 30 fps, and a 23-second reel renders in about 30 seconds on an M-series chip.
Who it is for. Creators of short explainers and developers who want to announce a skill or repository with a vertical video and no editor.
Good fit when
- You need a 20 to 30 second reel about a library or tool for social media
- A skill announcement for a repository should go out with a call-to-action keyword for comments
- You want a repeatable build with the script in `voice/script.json` and scenes in `src/scenes.js`
- You want to check the reel for text overlap and frame determinism before export with `reel check`
Not a fit when
- You need Linux or Windows: the instructions target macOS with Homebrew
- You need live footage with faces, actors or complex character acting: the style is flat with no facial animation
- The voice must sound different or use a language Kokoro does not cover
Example request
Make a paper-reel about uv, the Python package manager, with the call-to-action keyword UV and show the script before renderingLimitations
- It needs macOS, Node 18 or newer, ffmpeg and whisper.cpp. `reel doctor` checks every piece: the whisper model, a Python with `kokoro-onnx` and the Kokoro model files.
- Install downloads Chromium through Playwright and speech models, hundreds of megabytes in total.
- Render yields 1080 by 1920 at 30 fps, and the roughly 30-second time is quoted for Apple M-series chips.
- The style is fixed: paper shapes, a box mascot and stamps. A different visual language needs edits to the scenes in `src/scenes.js`.
- The project is independent and not affiliated with Anthropic. The renderer core and frame tools are vendored from `animate` by cth9191 (MIT).
How to disable. Delete `~/.claude/skills/paper-reel`. Finished reels live where you exported them, for example `~/reels/`, and are deleted by hand.
FAQ
How do I check that everything is installed?
Run `~/.claude/skills/paper-reel/bin/reel doctor`: it checks every piece and prints the exact command for anything missing.
Are API keys or accounts needed?
No. A local Kokoro makes the voice, whisper.cpp times the words, an HTML canvas with Playwright draws the frames and ffmpeg assembles them.
How do I run the pipeline by hand?
Create a project with `reel new ~/reels/my-reel`, edit `voice/script.json` and `src/scenes.js`, then run `reel voice`, `reel still`, `reel check` and `reel export`.
What size does the reel come out?
1080 by 1920 pixels at 30 fps, with a ducked lo-fi bed at about -14 LUFS.
What is the call-to-action keyword?
A word for comments, such as UV or REPO, that lands in the final scene. You give it in the request to Claude.
Discussion
No questions or reviews yet.
Related
SkillsA Claude Code skill that makes a Korean motion-graphics video from one request, from script to ElevenLabs voice, music and QA
A skill and CLI: the agent writes a script and scenes, while the voice, word timing and MP4 render run on your computer with no API keys
Nine skills and a Remotion library make short vertical explainer reels in a layered paper-cut style, from strategy to render.
Eleven Claude Code skills turn a talking-head clip into reels: titles, graphics behind you and split screen
A Claude Code skill for code-made video: it splits a clip into scenes and picks an engine for each, from Paper.js to Manim.
Claude Code skill: turns a one-line theme into a beat-synced HyperFrames showreel with morph transitions and generated sound.
README in short
The README shows an example reel and explains how the voice drives scenes. It continues with the macOS install, the request to Claude and the manual `reel` pipeline, links to the reference in `SKILL.md` and `references/style.md`, and a credits and MIT license section.