关于 OpenVideoGenerator MCP Server
面向 Kling、MiniMax、Luma 和 Hunyuan 的 AI 视频生成 MCP 服务,提供提示词蓝图
配置
远程服务无需本地安装。选择你使用的客户端,复制配置或使用一键安装。
https://openvideogenerator.com/api/mcpclaude mcp add --transport http zsthimself-open-video-generator https://openvideogenerator.com/api/mcp{"mcpServers":{"zsthimself-open-video-generator":{"type":"http","url":"https://openvideogenerator.com/api/mcp"}}}在 Cursor 中打开{"servers":{"zsthimself-open-video-generator":{"type":"http","url":"https://openvideogenerator.com/api/mcp"}}}在 VS Code 中打开{"mcpServers":{"zsthimself-open-video-generator":{"type":"http","url":"https://openvideogenerator.com/api/mcp"}}}{"mcpServers":{"zsthimself-open-video-generator":{"serverUrl":"https://openvideogenerator.com/api/mcp"}}}命令在终端中运行;JSON 粘贴到对应客户端的 MCP 配置后重启。需要授权的服务会在首次连接时提示登录。
https://openvideogenerator.com/api/mcp需要授权工具
README 原文English
来自项目 READMEREADME 暂无中文译文,以下为英文原文;中文介绍见上方。
OpenVideoGenerator MCP Server
Official Model Context Protocol (MCP) server and zero-dependency stdio bridge for OpenVideoGenerator — generate cinematic AI videos across Kling AI, MiniMax Hailuo, Luma Dream Machine, and Tencent Hunyuan Video directly from Cursor, Claude Desktop, Claude Code, and autonomous agents.
Available Tools
generate_video— Queue an asynchronous text-to-video or image-to-video render job and immediately receive ataskId.get_video_status— Poll render progress, stage status, and retrieve the final watermark-free 1080p MP4 URL.list_supported_engines— Inspect supported models, aspect ratios (16:9,9:16,1:1), durations, and credit costs.get_user_balance— Check remaining account credits and estimated generation capacity.search_prompt_blueprints— Search curated cinema-grade prompt blueprints with camera movements, lighting profiles, and negative prompts.
Option 1: Direct Streamable HTTP (Recommended)
Most modern MCP clients (Cursor, Claude Code, Antigravity) connect directly to the hosted Streamable HTTP endpoint:
- Endpoint:
https://openvideogenerator.com/api/mcp - API Key: Create a key at openvideogenerator.com/settings/apikeys (new accounts receive 50 free welcome credits).
{
"mcpServers": {
"open-video-generator": {
"url": "https://openvideogenerator.com/api/mcp",
"headers": {
"Authorization": "Bearer sk_YOUR_API_KEY_HERE"
}
}
}
}
Option 2: Local Stdio Bridge (npx)
For clients that require a local stdio command, run the zero-dependency Node.js 18+ bridge directly from GitHub:
{
"mcpServers": {
"open-video-generator": {
"command": "npx",
"args": ["-y", "github:zsthimself/open-video-generator-mcp"],
"env": {
"OPENVIDEO_API_KEY": "sk_YOUR_API_KEY_HERE"
}
}
}
}
Verify Connectivity
npx -y github:zsthimself/open-video-generator-mcp --check
Links
- Homepage: https://openvideogenerator.com
- Developer API Keys: https://openvideogenerator.com/settings/apikeys
- Prompt Library: https://openvideogenerator.com/prompts
- LLM Documentation: https://openvideogenerator.com/llms.txt
License
MIT
评论
一个开放协议,规定了 AI 应用如何以统一方式连接外部工具与数据源。客户端负责与模型对话,MCP 服务负责提供工具、资源和提示模板,两者通过标准消息通信,同一个服务可以被不同客户端复用。