VideoCutter

VideoCutter provides an MCP-enabled media processing platform that lets AI agents orchestrate video, audio, and image workflows through natural language and streamable MCP channels. It supports SSE for real-time progress updates and HTTP streamable mode for interactive workflows, enabling powerful, end-to-end automation of media projects.

How to use

You will connect your MCP client to VideoCutter’s MCP endpoints to drive media tasks. Use the SSE channel for ongoing progress updates and the HTTP Streamable channel when you need two-way interaction. Start by sending a simple task like generating an image from a text prompt, then progressively compose a multi-step workflow that includes video generation, overlaying text or graphics, and exporting final assets.

Key usage patterns include: (1) Natural language calling to trigger specific media processing actions, (2) Building intelligent workflows by chaining multiple tools, and (3) Receiving real-time progress and results as the task executes.

How to install

Prerequisites: ensure you have a supported runtime and network access to run MCP endpoints. You will typically need Node.js or a compatible runtime and a way to run local services.

Step-by-step setup you can follow locally:

# 1) Clone the project
git clone https://github.com/daimaxiuligong/VideoCutter.git

# 2) Enter the project directory
cd VideoCutter

# 3) Install dependencies
npm install

# 4) Build assets (if the project requires a build step)
npm run build

# 5) Start the MCP server (adjust if you use a different start script)
npm run start

Configuration and notes

VideoCutter exposes two MCP endpoints for integration. The SSE endpoint is for unidirectional streaming to report progress and status. The HTTP Streamable endpoint is for bidirectional streaming to enable interactive workflows with AI agents.

MCP endpoints to connect to: SSE at http://localhost:8000/mcp/sse and HTTP Streamable at http://localhost:8001/mcp/streamable. Use these URLs in your MCP client configuration to establish a connection and begin issuing requests.

Examples of MCP integration notes

• Natural language calls translate user intent into a sequence of media processing actions.

• AI agents can compose complex workflows by stitching together 67 professional tools for text, image, and video tasks.

Troubleshooting and tips

If you encounter connectivity issues, verify that the MCP endpoints are reachable from your client network and that any required authentication tokens are correctly provided.

Monitor progress via the SSE channel for long tasks and use the HTTP Streamable channel for interactive sessions that adjust in real time.

Available tools

text_generation

Generate high-quality text content from prompts using multiple AI models and languages.

image_generation

Create images from textual descriptions with adjustable resolution and styles.

video_generation

Produce short videos from descriptions or sequences, with configurable duration and resolution.

image_to_video

Convert static images into dynamic video sequences.

audio_enhancement

Improve vocal clarity and overall audio quality using advanced processing.

scene_detection

Automatically identify scene changes to assist precise editing.

speech_recognition

Extract subtitle text from audio tracks or video content.

text_to_speech

Synthesize speech from text with customizable voice profiles.

watermark_removal

Intelligently remove watermarks from images or frames when permitted.

overlay_batch

Apply multiple overlays (text or image) across assets in batch operations.

collage_creation

Create image collages or grid layouts for visual storytelling.

subtitle_generation

Automatically generate subtitles aligned with video content.

quality_optimization

AI-assisted optimization to balance video quality and file size.