Vidu MCP is an MCP server that provides access to Vidu's video generation models through applications supporting the Model Context Protocol, such as Claude Desktop and Cursor. It connects client interfaces directly to the Vidu cloud API, serving creative professionals, developers, and designers who need programmatic video generation capabilities inside their standard AI workspaces. By running the server via uvx, users can send prompt instructions directly to Vidu models to produce text-to-video, image-to-video, reference-to-video, and start-to-end video sequences. The server handles parameter inputs such as model version, duration, style, movement amplitude, aspect ratio, and resolution. When video generation completes, the system returns direct URL links to view, download, or share the generated outputs, removing the need to switch out of the client environment to execute visual generation workflows.
Category: Other & General Purpose
Tags: boilerplate, starter, template
uv/uvx package manager are installed on your system. 2. Obtain an API key from your Vidu Platform account settings. 3. Open your client configuration file, such as claude_desktop_config.json in Claude for Desktop or settings in Cursor. 4. Add the following entry to the mcpServers block: json { "mcpServers": { "Vidu": { "command": "uvx", "args": [ "vidu-mcp" ], "env": { "VIDU_API_KEY": "api-key-here", "VIDU_API_HOST": "api-host-here" } } } } 5. Replace api-key-here and api-host-here with your Vidu credentials. 6. Save the configuration file and restart or refresh your MCP client.Part of MCP Servers
Vidu MCP is a Python-based server implementing the Model Context Protocol to connect client applications with Vidu video generation APIs, enabling users to request text-to-video and image-to-video generation within their AI chat tools.
Install Python 3.10+ and the uv or uvx tool. Configure your MCP client settings file by adding Vidu under mcpServers with the command uvx, args vidu-mcp, and your VIDU_API_KEY and VIDU_API_HOST environment variables, then restart the client.
It is documented to work with any client supporting the Model Context Protocol, with specific configuration examples provided for Claude for Desktop and the Cursor editor.
Upon completion of a video generation task, the server delivers a URL link directly in the chat interface where you can view, download, and share the video file.
Video generation usually takes between 30 seconds and 5 minutes depending on the complexity of the prompt, model selection, current server load, and network conditions.