CapCutAPI/README.md

Connect AI generates via CapCutAPI Try it online

<div align="center">

👏👏👏👏 Celebrating 700 GitHub stars, giving away a 7000-point anonymous cloud rendering coupon: 08B88A2C-1D16-4CE1-982E-E3732F2655F3

</div>

Project Overview

CapCutAPI is a powerful cloud-based editing API that gives you precise control over AI-generated assets, including images, audio, video, and text. It provides the exact editing capabilities to assemble raw AI outputs, such as adjusting video speed or mirroring an image. This effectively solves the problem of a lack of precise control and difficulty in replicating AI-generated results, allowing you to easily transform creative ideas into polished videos. All these features are designed to be on par with the functionality of the Jianying software, ensuring you get a familiar and efficient editing experience in the cloud.

Core Advantages

1. Provides CapCut/Jianying-level editing capabilities through an API.

2. Allows real-time preview of editing results on a webpage, without the need for downloads, greatly facilitating workflow development.

3. Lets you download editing results and import them into CapCut/Jianying for secondary editing.

4. Enables you to use the API to generate videos from your editing results, achieving a fully cloud-based operation.

Demos

<div align="center">

MCP, create your own editing Agent

Combine AI-generated images and videos using tools

</div>

Key Features

Feature Module	API	MCP Protocol	Description
Draft Management	✅	✅	Create and save Jianying/CapCut draft files
Video Processing	✅	✅	Import, clip, transition, and apply effects to multiple video formats
Audio Editing	✅	✅	Audio tracks, volume control, sound effects processing
Image Processing	✅	✅	Image import, animation, masks, filters
Text Editing	✅	✅	Multi-style text, shadows, backgrounds, animations
Subtitle System	✅	✅	SRT subtitle import, style settings, time synchronization
Effects Engine	✅	✅	Visual effects, filters, transition animations
Sticker System	✅	✅	Sticker assets, position control, animation effects
Keyframes	✅	✅	Property animation, timeline control, easing functions
Media Analysis	✅	✅	Get video duration, detect format

Quick Start

1. System Requirements

• Python 3.10+
• Jianying or CapCut International version
• FFmpeg

2. Installation and Deployment

# 1. Clone the project
git clone https://github.com/sun-guannan/CapCutAPI.git
cd CapCutAPI

# 2. Create a virtual environment (recommended)
python -m venv venv-capcut
source venv-capcut/bin/activate  # Linux/macOS
# or venv-capcut\Scripts\activate  # Windows

# 3. Install dependencies
pip install -r requirements.txt      # HTTP API basic dependencies
pip install -r requirements-mcp.txt  # MCP protocol support (optional)

# 4. Configuration file
cp config.json.example config.json
# Edit config.json as needed

3. Start the service

python capcut_server.py # Start the HTTP API server, default port: 9001

python mcp_server.py # Start the MCP protocol service, supports stdio communication

MCP Integration Guide

MCP Documentation in Chinese • MCP English Guide

1. Client Configuration

Create or update the mcp_config.json configuration file:

{
  "mcpServers": {
    "capcut-api": {
      "command": "python3",
      "args": ["mcp_server.py"],
      "cwd": "/path/to/CapCutAPI",
      "env": {
        "PYTHONPATH": "/path/to/CapCutAPI",
        "DEBUG": "0"
      }
    }
  }
}

2. Connection Test

# Test MCP connection
python test_mcp_client.py

# Expected output
✅ MCP server started successfully
✅ Got 11 available tools
✅ Draft creation test passed

Usage Examples

1. API Example

Add video material

import requests

# Add background video
response = requests.post("http://localhost:9001/add_video", json={
    "video_url": "https://example.com/background.mp4",
    "start": 0,
    "end": 10
    "volume": 0.8,
    "transition": "fade_in"
})

print(f"Video addition result: {response.json()}")

Create stylized text

import requests

# Add title text
response = requests.post("http://localhost:9001/add_text", json={
    "text": "Welcome to CapCutAPI",
    "start": 0,
    "end": 5,
    "font": "Source Han Sans",
    "font_color": "#FFD700",
    "font_size": 48,
    "shadow_enabled": True,
    "background_color": "#000000"
})

print(f"Text addition result: {response.json()}")

More examples can be found in the example.py file.

2. MCP Protocol Example

Complete workflow

# 1. Create a new project
draft = mcp_client.call_tool("create_draft", {
    "width": 1080,
    "height": 1920
})
draft_id = draft["result"]["draft_id"]

# 2. Add background video
mcp_client.call_tool("add_video", {
    "video_url": "https://example.com/bg.mp4",
    "draft_id": draft_id,
    "start": 0,
    "end": 10,
    "volume": 0.6
})

# 3. Add title text
mcp_client.call_tool("add_text", {
    "text": "AI-Driven Video Production",
    "draft_id": draft_id,
    "start": 1,
    "end": 6,
    "font_size": 56,
    "shadow_enabled": True,
    "background_color": "#1E1E1E"
})

# 4. Add keyframe animation
mcp_client.call_tool("add_video_keyframe", {
    "draft_id": draft_id,
    "track_name": "main",
    "property_types": ["scale_x", "scale_y", "alpha"],
    "times": [0, 2, 4],
    "values": ["1.0", "1.2", "0.8"]
})

# 5. Save the project
result = mcp_client.call_tool("save_draft", {
    "draft_id": draft_id
})

print(f"Project saved: {result['result']['draft_url']}")

Advanced text effects

# Multi-style colored text
mcp_client.call_tool("add_text", {
    "text": "Colored text effect demonstration",
    "draft_id": draft_id,
    "start": 2,
    "end": 8,
    "font_size": 42,
    "shadow_enabled": True,
    "shadow_color": "#FFFFFF",
    "background_alpha": 0.8,
    "background_round_radius": 20,
    "text_styles": [
        {"start": 0, "end": 2, "font_color": "#FF6B6B"},
        {"start": 2, "end": 4, "font_color": "#4ECDC4"},
        {"start": 4, "end": 6, "font_color": "#45B7D1"}
    ]
})

3. Downloading Drafts

Calling save_draft will generate a folder starting with dfd_ in the current directory of capcut_server.py. Copy this to the CapCut/Jianying drafts directory to see the generated draft in the application.

Community & Support

We welcome contributions of all forms! Our iteration rules are:

• No direct PRs to main
• PRs can be submitted to the dev branch
• Merges from dev to main and releases will happen every Monday

Join our group

• Feedback on issues
• Feature suggestions
• Latest news

🤝 Collaboration Opportunities

• Overseas Video Production: Want to use this API for batch production of overseas videos? I offer free consulting services to help you use this API. In return, I'll ask for the production workflow template to be open-sourced in the template directory of this project.

• Join us: Our goal is to provide a stable and reliable video editing tool that integrates well with AI-generated images, videos, and audio. If you are interested, you can start by translating the Chinese content in the project into English! Submit a PR and I'll see it. For more in-depth involvement, the code for the MCP Editing Agent, web-based editing client, and cloud rendering modules has not been open-sourced yet.

• Contact:

  • WeChat: sguann
  • Douyin: 剪映草稿助手

📈 Star History

<div align="center">

</div>

Made with ❤️ by the CapCutAPI Community