Content
# Tool List
> **Complete tutorial on operating ChatCut with OpenClaw** - Let AI help you edit videos, from configuring MCP to generating Motion Graphic
[](https://opensource.org/licenses/MIT)
[](https://chatcut.io)
[](https://openclaw.ai)
[](./docs/03-tool-catalog.md)
---
## What is this?
**ChatCut** is an AI-driven video editor, and **OpenClaw** is an AI Agent runtime. They are connected through the **Model Context Protocol (MCP)**, allowing AI to assist with video editing.
This project is a **complete practical tutorial**, teaching you from scratch:
- 🛠️ **Configure ChatCut MCP** - 5-minute setup for OpenClaw ↔ ChatCut
- 🤖 **Operate video with AI Agent** - 53 tools at your fingertips
- 🎨 **Create Motion Graphic with JSX** - Directly generate cinematic effects
- 🚨 **Pitfall guide** - Avoid 10+ common pitfalls
## ⚡ Quick Start
### 1. Install OpenClaw
```bash
# macOS
brew install openclaw
# or download from https://openclaw.ai/download
```
### 2. Register ChatCut and configure MCP
```bash
# Use OpenClaw to add ChatCut MCP server
openclaw mcp add chatcut \
--url https://api.chatcut.io/api/external-mcp/mcp \
--auth oauth --transport streamable-http
# Login (opens browser for authorization)
openclaw mcp login chatcut
```
### 3. Let AI list your video projects
In the OpenClaw dialogue box, say:
> "List my projects with chatcut"
OpenClaw will automatically call the `chatcut.list_projects` tool to provide your project list.
🎉 **That's it! You've connected to ChatCut!**
---
## 🎯 Core Capabilities
| Category | Number of Tools | Typical Usage |
|---|---|---|
| 📁 Project Management | 8 | Create/copy/delete/restore projects |
| 🎬 Timeline Editing | 7 | Edit tracks, cut, move, mark |
| 📥 Media Import | 5 | Local files / URL / S3 presigned |
| 🎨 Motion Graphic | 4 | **JSX generates cinematic effects** |
| 🤖 AI Generation | 5 | Video/music/voiceover/sound effects/shaders |
| 📝 Subtitle Transcription | 7 | Automatic subtitle generation + filler word removal |
| 🛒 Media Search | 2 | Search Pexels/Pixabay public media |
| 📤 Export | 2 | h264/vp8/FCP XML/ProRes |
| 👁️ Preview Analysis | 3 | Timeline screenshot rendering |
| ⚙️ Task Tracking | 2 | Asynchronous AI task status |
**Complete Tool Catalog** → [`docs/03-tool-catalog.md`](./docs/03-tool-catalog.md)
---
## 📚 Tutorial Documents
| No. | Topic | What You'll Learn |
|---|---|---|
| 01 | [Quickstart](./docs/01-quickstart.md) | 5-minute ChatCut MCP setup |
| 02 | [MCP Setup Details](./docs/02-mcp-setup.md) | 4 connection methods + OAuth renewal |
| 03 | [Tool Catalog](./docs/03-tool-catalog.md) | Complete list of 53 tools |
| 04 | [Pitfall Summary](./docs/04-pitfalls.md) | 10+ real pitfalls + solutions |
| 05 | [Practical Recipes](./docs/05-recipes.md) | 5 end-to-end scenarios |
| 06 | [Motion Graphic Advanced](./docs/06-motion-graphics.md) | JSX for cinematic effects |
---
## 🚀 Practical Examples
```bash
cd examples
# List your projects
node 01-list-projects.mjs
# Read timeline
node 02-read-timeline.mjs
# Generate Motion Graphic with JSX (killer scenario!)
node 03-create-motion-graphic.mjs
# AI voiceover
python 04-generate-voice.py
# Search public media
node 06-stock-search.mjs "sunset beach"
# End-to-end pipeline
node 08-full-pipeline.mjs
```
---
## 🌟 Killer Scenarios
### Scenario 1: Generate a movie intro in one sentence
> "Create a 5-second movie intro with ChatCut, black background with golden text"
AI will automatically:
1. Create a project
2. Write a Motion Graphic with JSX
3. Add to timeline
4. Export preview
### Scenario 2: Automatic cleanup of interview footage
> "Clean up this 60-minute interview footage, removing 'um' and 'ah' and compressing silence"
AI will automatically:
1. Upload video
2. Call `clean_script` to remove filler words
3. Compress silent intervals
4. Regenerate subtitles
5. Export
### Scenario 3: One-click B-Roll replacement
> "Add 5 B-Roll segments of solar panels, wind turbines, and forests to my 'carbon neutrality' video"
AI will automatically:
1. Search for solar panels/wind turbines/forests in stock media
2. Download to media pool
3. Insert between VO segments on timeline
4. Adjust duration automatically
5. Export
---
## 🛡️ Security & Privacy
- **Don't submit OAuth tokens to Git** - `.gitignore` ignores `.openclaw/` and `chatcut-*.json` by default
- **Least privilege principle** - OpenClaw exposes only necessary tools through MCP
- **Local priority** - All video processing is done in ChatCut cloud, token used only for authentication
---
## 🤝 Contribution
PRs welcome! Especially:
- New example scenarios
- More pitfall cases
- Translations (English / 日本語 / Español)
- Bug fixes
Please open an issue to discuss major changes.
---
## 📄 License
MIT © 2026 chonpszhou
---
## 🙏 Acknowledgements
- [ChatCut](https://chatcut.io) - AI video editor
- [OpenClaw](https://openclaw.ai) - AI Agent runtime
- [Model Context Protocol](https://modelcontextprotocol.io) - MCP protocol
---
## ⭐ Star History
If this project helps you avoid pitfalls, star it ⭐ - that's the biggest motivation for continuous updates!
