ShotFlow MCP Server — AI分镜脚本与视频生成智能体,支持生成专业分镜、视频、镜头重生成、配额查询等能力。
让 AI Agent 直接调用专业分镜脚本生成能力
ShotFlow MCP Server 是 ShotFlow 的 Model Context Protocol 接口实现,让任何支持 MCP 的 AI 客户端(Claude Desktop、Cursor、Cline、VS Code Copilot 等)都能直接调用专业分镜脚本与视频生成能力。
无需安装,直接配置 URL 即可使用:
https://aishotflow.cn/mcp
| Tool | 说明 |
|---|---|
generate_storyboard | 从剧本/创意描述生成专业分镜脚本(6种创作类型 × 7种视觉风格) |
generate_video | 从分镜镜头生成视频(Pro 功能) |
check_video_status | 查询视频生成状态 |
regenerate_shot | 单镜头重新生成 |
list_video_tasks | 列出视频任务历史 |
check_quota | 查询今日配额 |
check_health | 检查服务健康状态 |
| Resource | 说明 |
|---|---|
shotflow://templates/creative-types | 6种创作类型模板与默认参数 |
shotflow://templates/visual-styles | 7种视觉风格参考与关键词 |
编辑配置文件(macOS: ~/Library/Application Support/Claude/claude_desktop_config.json,Windows: %APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"shotflow": {
"url": "https://aishotflow.cn/mcp"
}
}
}
重启 Claude Desktop,点击「Attached tools」确认 shotflow 已加载。
在项目根目录创建 .cursor/mcp.json:
{
"mcpServers": {
"shotflow": {
"url": "https://aishotflow.cn/mcp"
}
}
}
在项目根目录创建 .vscode/mcp.json:
{
"servers": {
"shotflow": {
"type": "http",
"url": "https://aishotflow.cn/mcp"
}
}
}
编辑 ~/.cline/mcp_settings.json:
{
"mcpServers": {
"shotflow": {
"url": "https://aishotflow.cn/mcp"
}
}
}
告诉你的 AI 客户端:
用 shotflow 的 generate_storyboard 工具,帮我生成一个短视频分镜:一个创业者熬夜做产品的故事,3个场景,温情风格,9:16竖屏
AI 会自动调用 MCP Tool,返回结构化分镜结果。
读取 shotflow 的 creative-types 资源,告诉我有哪些创作类型可选
其他 AI Agent 可通过以下 URL 自动发现 ShotFlow 的能力:
https://aishotflow.cn/mcp/agent-card
localhost:3003git clone https://github.com/YOUR_USERNAME/shotflow-mcp.git
cd shotflow-mcp
# 无需 npm install,零依赖
# 开发模式
node shotflow-mcp-server.js
# 默认端口 3004
# 生产部署(systemd)
sudo cp shotflow-mcp.service /etc/systemd/system/
sudo systemctl enable shotflow-mcp
sudo systemctl start shotflow-mcp
| 变量 | 默认值 | 说明 |
|---|---|---|
MCP_PORT | 3004 | MCP Server 监听端口 |
SHOTFLOW_API_URL | http://localhost:3003 | ShotFlow Express 地址 |
NODE_ENV | development | 环境模式 |
AI Client (Claude/Cursor/...)
↓ Streamable HTTP (JSON-RPC 2.0)
Nginx :443
└── /mcp → ShotFlow MCP Server :3004
└── HTTP → ShotFlow Express :3003
proxy_buffering off 配合 Nginx 实现流式进度推送shotflow-mcp/
├── shotflow-mcp-server.js # MCP Server 主文件
├── tool-definitions.js # 7个 Tool 的 JSON Schema 定义
├── resource-definitions.js # 2个 Resource 定义
├── prompt-builder.js # Prompt 构造与模型配置解析
├── package.json # npm 包配置
├── shotflow-mcp.service # systemd 服务文件
├── INTEGRATION.md # 详细接入文档
├── README.md # 本文件
└── test/
└── test-mcp.js # 测试套件
# 启动本地服务
node shotflow-mcp-server.js &
# 运行测试
node test/test-mcp.js
MIT License — 自由使用、修改和分发。
Source-derived launch command. Check the maintainer’s required arguments and credentials before running:
npx -y shotflow-mcp-serverMerge this template into ~/Library/Application Support/Claude/claude_desktop_config.json. Keep existing servers. Add any arguments, credentials, and permissions required by the maintainer; this template has not been install-tested.
{
"mcpServers": {
"io-github-johnzhang-ui-shotflow-mcp": {
"command": "npx",
"args": [
"-y",
"shotflow-mcp-server"
]
}
}
}Restart Claude Desktop completely for changes to take effect. Confirm the server appears connected in the client’s tool list, then try a read-only example from its documentation.
Claude Desktop setup referenceshotflow-mcp-servernpmio.github.JOHNZhang-ui/shotflow-mcp works with any MCP-compatible client. Copy the config snippet from the Configuration section above and add it to the file shown for your client, then restart the application.
~/Library/Application Support/Claude/claude_desktop_config.jsonRestart Claude Desktop completely for changes to take effect.~/.cursor/mcp.jsonRestart Cursor for changes to take effect..vscode/mcp.jsonReload VS Code window for changes to take effect.~/.codeium/windsurf/mcp_config.jsonRestart Windsurf for changes to take effect..mcp.jsonSave at the project root, then start Claude Code in that project and review the MCP server approval prompt. Keep real credentials out of shared files.