AIEV workflow
Claude Code directs a video: transcription, script, HyperFrames scenes, subtitles and Remotion assembly with a localhost dashboard
- 157 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Claude Code runs commands and writes project files in the working folder during editing
- The tunnel script exposes the local server externally for phone uploads, and paid Gemini calls need a key
Install
- Works in
Manual install
Clone `https://github.com/notivn/AIEV.git`, enter the folder and run `./start/start.sh` (on Windows `start\start.bat`, on macOS `start/start.command`). Read the environment check output and do not start installation without the user agreement. The user must sign in to Claude and add the Gemini key themselves. The dashboard opens at http://localhost:6868.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Claude Code runs commands and writes project files in the working folder during editing
- The tunnel script exposes the local server externally for phone uploads, and paid Gemini calls need a key
This is third-party code. Review the repository files before installing.
What it does
AIEV ("AI Edit Video" by noti.vn) is a local automatic video editing system where Claude acts as director. You upload a clip, briefly describe the task and press "Start editing with AI". Claude Code then transcribes the audio, writes an editing script, builds kinetic typography scenes with HyperFrames (HTML and GSAP), adds karaoke subtitles, beat-synced zooms and sound effects, assembles a timeline in Remotion and exports an MP4. The whole process is visible in the web dashboard at `http://localhost:6868`. Editing runs in two passes, draft and final, with every frame verified. There is a style design (sets of colors, fonts, logos and tone; fonts download from Google Fonts), color grading with preview (14 presets and manual tuning), a key layout with main and related keys in the upper and lower bands, background music and AI illustrations: Claude picks moments and Gemini draws pictures in the chosen style (about 0.05 dollars per image). The skill set lives in `.claude/skills/`: `auto-cut`, `color-grading`, `background-music`, `gsap`, `hyperframes`, `hyperframes-cli`, `hyperframes-registry`, `ai-illustrations`. It starts with `start/` scripts for Windows, macOS and Linux that check the environment and can install what is missing.
Who it is for. Short video creators, editors and studios, Vietnamese and international, who need automatic clip editing with subtitles and graphics without manual timeline work.
Good fit when
- A source clip must become a dynamic video with karaoke-style subtitles and brand graphics
- You need several brand style kits and a consistent look for every episode
- The video needs AI illustrations at the right moments per the script
- You want process control through a web dashboard and a draft before the final render
Not a fit when
- You need a classic video editor for precise manual track work: AIEV builds the edit automatically
- You have no Claude Code subscription or login: the system uses Claude as director
- You need editing without Node.js 22 and FFmpeg: both are required for assembly
Example request
Edit my clip: cut the pauses, add karaoke subtitles and three typography scenes in my brand styleLimitations
You need Node.js 22 or newer, FFmpeg, Google Chrome and a logged-in Claude Code. The `start/doctor.mjs` script checks the environment and can install what is missing, while you add the Claude login and Gemini key yourself. faster-whisper is recommended for subtitles and an NVENC GPU or Apple Silicon for render speed.
The Claude Code agent works with your files and runs commands, so give it only a working folder. AI illustrations spend paid Gemini calls. Subtitles are tuned for Vietnamese, and other languages depend on the recognition model. Phone uploads use the `tunnel` script, which exposes the local server externally, so enable it deliberately.
How to disable. Stop the system with `start/stop.sh` (or `stop.bat`, `stop.command`) and delete the AIEV folder. The skills live in `.claude/skills/` inside the repository folder and need no separate removal.
FAQ
How do I start AIEV?
Clone the repository and run the script from the `start/` folder: `start.bat` on Windows, `start.command` on macOS, `./start/start.sh` on Linux. Then open http://localhost:6868.
What do I need to install first?
Node.js 22 or newer, FFmpeg, Google Chrome and a logged-in Claude Code. The start script checks the environment and offers to install what is missing.
Do I need a Gemini key?
Only for AI illustrations and backgrounds. Editing, subtitles and HyperFrames scenes work without it, and an image costs about 0.05 dollars.
Which subtitles does AIEV make?
Karaoke subtitles with keyword highlighting based on faster-whisper. The author specifically fixed issues with Vietnamese diacritics.
Can I use my own skills?
Yes, `.claude/skills/` is an ordinary folder where you can add your own. Base skills include hyperframes, gsap, auto-cut and color-grading.
Discussion
No questions or reviews yet.
Related
WorkflowsEleven Claude Code skills turn a talking-head clip into reels: titles, graphics behind you and split screen
A video editing skill for Claude Code and Codex: the agent cuts pauses and filler words, grades color, burns subtitles and adds animated overlays
Skill edits video from a form brief: cuts, logo, on-screen inserts, zoom and music, all running locally.
Skills, a CLI and an MCP server that let Claude Code or Codex make video from a brief: editing, graphics, voiceover, subtitles
A Claude Code skill that builds motion graphics in code: 21 ready looks, five questions, three approval stills and an MP4 render.
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.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| AIEV (this card) | no | 157 | |
| HyperFrames | no | 83,283 | a framework for HTML video, while AIEV adds automatic direction and a dashboard |
| YouTube AI Video Editing System | no | 145 | a YouTube video editing system without HyperFrames subtitles |
README in short
The repository describes an automatic editing pipeline where Claude drives HyperFrames and Remotion while a person observes through a web dashboard. It has English and Vietnamese READMEs, start and update scripts and a skill set for scene generation.
README badge
Original SKILL.md
--- name: ai-illustrations description: Generate illustrations with Gemini and composite them into the video being edited - how to pick the moments that need an image, write the prompt, call the /api/illustrations endpoint (images MUST match the selected Style Design), and composite into the composition. Read when the brief enables "Ảnh minh họa AI" (AI illustrations) or the user asks for generated illustrations for a video. --- # AI Illustrations - Claude directs, Gemini draws, brand stays consistent ## ⛔ GATE - read before anything else This skill applies **only when the edit prompt says "Ảnh minh họa AI: BẬT"** (AI illustrations ON). If the edit prompt says **"Ảnh minh họa AI: TẮT"**, or has no illustration line at all: **generate nothing.** Do not call `POST /api/illustrations`, do not composite AI-generated images into any scene, and do not let another skill's "use an illustration here" suggestion talk you into it. Build the visuals from what the project already has: assets in `assets/`, brand logos, and HyperFrames scenes (typography, graphics, shapes, motion). The server enforces this too - with the switch off the endpoint returns `409 ILLUSTRATIONS_DISABLED` before it spends anything. **Why this section exists:** the edit prompt used to stay silent when the switch was off, this skill got read anyway, and finished videos came back full of Gemini images the user had explicitly turned off. A silent prompt is not permission. ## Directing principles 1. **Pick moments deliberately, do not sprinkle images everywhere.** Read the transcript/video content and choose 2-5 points where an illustration explains more than the words do: abstract concepts, key numbers, a product/setting being mentioned, before-vs-after comparisons. DO NOT illustrate talking-head passages that already hold attention on their own. 2. **Each image costs ~$0.05-0.07** - treat it as real money. A short video (<60s) usually needs only 2-4 images. 3. **Illustrations are BACKGROUND visuals - no text by default.** The server already bans text in the prompt; numbers and labels are placed on top by HyperFrames/Remotion (correct font + brand color, no typos). The only exception: the brief enables "Ảnh có chữ" (`illustrationText`) or the edit prompt says "ĐƯỢC PHÉP CÓ CHỮ" (text allowed) -> see the "Images with text (allowText)" section below. 4. **Style Design is LAW - follow it 100%, no exceptions.** Every illustration must match the project's selected style: ALWAY