Crabcut
@realcrabcut
Crabcut について
Turn YouTube videos into short clips — from Claude, Cursor, or any AI assistant that supports MCP.
基本情報
設定
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"crabcut": {
"command": "npx",
"args": [
"@crabcut/mcp-server"
],
"env": {
"CRABCUT_API_KEY": "<YOUR_API_KEY>"
}
}
}
}ツール
5Start clip generation from a YouTube URL. Returns a `project_id` right away.
Check progress and get clips when ready. Clips are sorted by score (best first).
List your projects with status and clip counts.
Get full details for a single clip.
See your plan, remaining credits, and usage.
概要
What is Crabcut?
Crabcut is an MCP server that turns YouTube videos into short, vertical (9:16) clips with subtitles and download links. It is designed for AI assistants like Claude, Cursor, or any MCP‑compatible automation workflow.
How to use Crabcut?
Get an API key at app.crabcut.ai/developers, then run CRABCUT_API_KEY=sk_live_... npx @crabcut/mcp-server. The server connects via stdio; for remote usage use the endpoint https://mcp.crabcut.ai/mcp with Bearer token authentication.
Key features of Crabcut
- Generates vertical video clips from any YouTube URL
- Automatically adds subtitles to every clip
- Returns clips sorted by an AI‑generated quality score
- Provides both polling and webhook callback workflows
- Offers tools to list projects and check credit usage
- Supports streamable HTTP and stdio transports
Use cases of Crabcut
- Automatically create short‑form social media clips from long YouTube videos
- Extract the best moments of a podcast or talk for highlights
- Generate multiple clip options and pick the highest‑scoring one
- Integrate clip creation into no‑code automation pipelines like n8n
FAQ from Crabcut
How does Crabcut differ from manual clip cutting?
Crabcut uses AI to detect the best moments, reframes them to vertical format, and adds subtitles automatically — all from a single YouTube link and an API call.
What runtime or API key is required?
You need a Crabcut API key (free/paid plan) and Node.js to run the npx command. No additional databases or local storage are required — all processing happens on Crabcut’s servers.
Where are the generated clips stored?
Clips are hosted on Crabcut’s CDN and accessible via download_url in the response. They are not stored locally; you can re‑download them as long as the URL is valid.
What transports and authentication are supported?
The server supports stdio (for local MCP clients) and Streamable HTTP (for remote clients like n8n). Authentication is via an API key passed as an environment variable or Bearer token.
Are there any rate limits or known error conditions?
Errors include invalid YouTube URL (400), unauthorized API key (401), insufficient credits (402), rate limit exceeded (429), and internal server errors (500). Credits are consumed per project.
「AI とエージェント」の他のコンテンツ
Intervals.icu MCP Server
mvilanovaModel Context Protocol (MCP) server for connecting Claude and ChatGPT with the Intervals.icu API.
Open Multi-Agent Canvas
CopilotKitThe open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add MCP servers for deep research
MCP Server - Remote MacOs Use
baryhuangThe only general AI agent that does NOT requires extra API key, giving you full control on your local and remote MacOs from Claude Desktop App
Gemini MCP Server
aliargunMCP server implementation for Google's Gemini API
1MCP - One MCP Server for All
1mcp-appA unified Model Context Protocol server implementation that aggregates multiple MCP servers into one.
コメント