claude-youtube-editor skill
Claude Code skills that turn a talking-head recording into a YouTube video with graphics, sound and thumbnails
- 357 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Scripts upload videos to YouTube from your account through OAuth
- The pipeline uses AssemblyAI, ElevenLabs and Gemini keys from the .env file and sends audio and frames to those services
- The agent runs Python, Node and ffmpeg inside the project folder
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/claude-youtube-editorDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Scripts upload videos to YouTube from your account through OAuth
- The pipeline uses AssemblyAI, ElevenLabs and Gemini keys from the .env file and sends audio and frames to those services
- The agent runs Python, Node and ffmpeg inside the project folder
This is third-party code. Review the repository files before installing.
What it does
claude-youtube-editor is a set of eight Claude Code skills and Python tools that take a raw talking-head recording to a finished video. You record only yourself. Claude builds diagrams, browser and terminal mockups, title cards and transitions as Remotion code and composites them over your cut. The work runs in six steps. The `/clean-cut` skill transcribes the recording through AssemblyAI and removes the unwanted parts according to your `cuts.json` file. `/make-tsx` and `/fake-screencast` draw overlays timed to the words in the transcript. `/clean-audio` cleans the voice, `/suggest-sfx` places sound effects, and `/packaging` proposes one title and three thumbnail variants for YouTube's A/B test. The last step, the `tools/yt_upload.py` script, checks the video for ghost speech and audio-video drift, then uploads it to YouTube as a private draft. The `/brand-setup` skill changes the palette, fonts and wordmark to fit your channel, because the project ships with the author's own look.
Who it is for. For YouTube creators and tutorial authors who film themselves on camera and want a finished video without manual editing.
Good fit when
- You recorded a video with yourself on camera and want a finished video with overlays timed to the words
- You need diagrams and terminal or browser mockups built as code, with no screen recording
- You want sound effects placed on exact words and the voice levels cleaned up
- Before publishing you need a title and three thumbnail variants for YouTube's A/B test
- You need a pipeline that checks a video for audio drift and uploads it as a private draft
Not a fit when
- You need a script for the video: the kit has no script-writing skill
- You have no source recording: the repository holds only finished example overlays, no raw footage
- Thumbnails need photos of your face: you have to put them into media/library/faces yourself
- You want a timeline editor: the project works through a dialogue with Claude Code
- Your project or client rules forbid sending the audio and frames to outside services
Example request
Run clean cut on videos/video-1, then add TSX beats and suggest sound effectsLimitations
You need Claude Code, Python 3.10 or newer, Node 18 or newer, plus ffmpeg and ffprobe on PATH. The project itself is free under MIT, but the services cost money: AssemblyAI for transcription (the free tier is enough to try it), ElevenLabs for voice isolation, effects and music, Gemini for thumbnails. Voice cleaning also has a local RNNoise method that needs no key. Uploading to YouTube goes through OAuth.
The `/packaging` rules are calibrated on the author's channel data. The author suggests building your own CTR numbers after roughly ten videos. Transitions, zoom effects and color filters are on the roadmap, with no ready library yet.
How to disable. The skills live in the .claude/skills folder of the cloned repository, so deleting the project folder or opening Claude Code in another directory is enough. You can revoke YouTube access in your Google account settings and remove the keys from the .env file.
FAQ
How much does claude-youtube-editor cost to use?
The kit itself is free under the MIT license. You pay for Claude Code, ElevenLabs (voice, effects, music) and Gemini (thumbnails). AssemblyAI has a free tier to try, and the voice can be cleaned with the local RNNoise method without a key.
Do I need to record my screen for the overlays?
No, every screen moment is built as Remotion code, including browser and VS Code mockups. The repository ships 37 finished shots from a published video, and you can open them with `npm run studio` in the remotion folder without any source footage.
How do I stop my videos from looking like the repository author's?
Run `/brand-setup` once. The skill interviews you about palette, fonts and wordmark, rewrites brand.md, brand.ts and fonts.ts, and renders a proof card.
Where does the script upload the finished video?
To YouTube as a private draft through OAuth. The tools/yt_upload_SETUP.md file explains the access setup.
What is the vidtsx-2d-generator skill for?
It holds the low-level rules that keep a Remotion shot from crashing during render. The other skills in the kit, such as `/make-tsx`, build on those rules.
Discussion
No questions or reviews yet.
Related
SkillsClaude Code Video Toolkit
claude-code-video-toolkit
A video production kit for Claude Code: script, voiceover, music, visuals and MP4 rendering through Remotion, with some steps on paid cloud services
Five Claude Code skills that produce faceless vertical YouTube Shorts from Remotion code, video models or paper collage
A Claude Code skill that downloads a YouTube video, splits it into semantic chapters, cuts chosen clips, translates subtitles and burns bilingual ones in
Claude Motion Design
claude-motion-design
A Claude Code skill for motion videos in pure code: HTML, Playwright and ffmpeg without After Effects and Remotion. Chromium, ffmpeg and Python needed
A video editing skill for Claude Code and Codex: the agent cuts pauses and filler words, grades color, burns subtitles and adds animated overlays
Skills, a CLI and an MCP server that let Claude Code or Codex make video from a brief: editing, graphics, voiceover, subtitles
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| claude-youtube-editor (this entry) | no | 357 | |
| Claude Code Video Toolkit | no | 2,378 | a kit for building video from scratch: script, voiceover, music, visuals |
| OpenMontage | no | 72,968 | an agentic video studio that writes the script and voices the video itself |
| YouTube agent skill | no | 3,604 | skills for the script, title and thumbnail, with no editing of the recording |
README in short
The project's main claim is that you record yourself and Claude Code does the rest. A separate list states plainly what the repository lacks: a script writer, source footage, a face kit and CTR data.
README badge
Original SKILL.md
--- name: brand-setup description: Makes this repo's videos look like YOUR channel instead of the house default — interviews you for palette, fonts, wordmark, motion energy, delivery specs and SFX taste, then rewrites brand.md + remotion/src/brand.ts + remotion/src/fonts.ts together and renders a proof card so you SEE it. Use whenever the user wants to set up or change their brand, rebrand the repo, "make it my colors/fonts", change the accent color, set their wordmark, adjust the motion feel, or asks why their videos look like someone else's. Run it once before the first video. Not for per-video art direction (that is /make-tsx) and not for thumbnails, which are deliberately louder than the brand (that is /packaging). --- # Brand Setup **Run this once, before your first video.** This repo ships with a house brand — a calm, premium indigo look. It's a real, working brand, not a placeholder, and you're welcome to keep it. But if you never touch it, every video you make here looks like the channel it came from. This skill fixes that in one pass. ## What this skill owns Three files that **must stay in sync**, and this skill is the only thing that should write all three: | File | What it holds | Who reads it | |---|---|---| | `brand.md` | the contract, in prose | every step-2+ skill (`/make-tsx`, `/suggest-sfx`, `/packaging`) | | `remotion/src/brand.ts` | the same tokens as code — `COLORS`, `GRADIENT`, `RADIUS`, `SHADOW`, `EASINGS`, `BRAND` | every shot | | `remotion/src/fonts.ts` | the font families, loaded from `@remotion/google-fonts` | every shot | **They drift silently.** `brand.md` says indigo, `brand.ts` says teal, and nothing errors — the docs just quietly stop describing the videos. That's why one skill writes all three in one pass. Every shot imports from `brand.ts` / `fonts.ts`, so changing a value here re-brands **every video in the repo, retroactively.** That's the whole point of a contract. ## The flow --- ### Stage 1 — Interview (ONE message, not twenty questions) Ask it all at once, and **say up front that "keep the default" is a valid answer to any of it** — most people care about two or three of these and have no opinion on the rest. 1. **Channel identity** — the wordmark text, and what the channel is about in a sentence. The wordmark renders in three parts with the **middle part in the accent color** (`['Learn','With','Hasan']` → the "With" is indigo). Two-part marks use an empty third (`['Acme','Labs','']`). Also: the sign-off line for the end card. 2. **The feeling** — calm/premium, loud/energetic, playful, editorial, technical? This drives motion and shape more than color does. 3. **Palette** — do they have brand colors already (paste hex), or should you propose from the feeling? Light base or dark base? The accent is the single most important choice: it's the key words, the highlights, the CTAs, the wordmark's middle. 4. **Type** — existing brand fonts, or propose to fit the feeling? Must be Google Fonts (see the hard gate in Stage 3).