Wavespeed

Provides a standardized MCP interface to WaveSpeed AI for image and video generation tasks via configurable environment, CLI, and config-file options.

python

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "wavespeedai-mcp-server": {
      "command": "wavespeed-mcp",
      "args": [],
      "env": {
        "WAVESPEED_API_KEY": "your-api-key-here",
        "WAVESPEED_API_HOST": "https://api.wavespeed.ai",
        "WAVESPEED_LOG_FILE": "/tmp/wavespeed-mcp.log",
        "WAVESPEED_LOG_LEVEL": "INFO",
        "WAVESPEED_MCP_BASE_PATH": "~/Desktop",
        "WAVESPEED_API_RESOURCE_MODE": "url",
        "WAVESPEED_API_VIDEO_ENDPOINT": "/wavespeed-ai/wan-2.1/i2v-480p-lora",
        "WAVESPEED_API_TEXT_TO_IMAGE_ENDPOINT": "/wavespeed-ai/flux-dev",
        "WAVESPEED_API_IMAGE_TO_IMAGE_ENDPOINT": "/wavespeed-ai/flux-kontext-pro"
      }
    }
  }
}

WavespeedMCP exposes WaveSpeed AI’s image and video generation capabilities through a standardized MCP interface, letting you run the MCP server locally or connect to it from your client to manage jobs, monitor progress, and retrieve outputs with flexible resource handling and robust error signaling.

How to use

You run WavespeedMCP as a local or remote MCP server and connect to it from an MCP client. Start the server with your API key, then use your client to submit text-to-image, image-to-image, or video generation tasks. The server handles job submission, polling for progress, and returning outputs in your preferred resource mode (URL, local file, or base64). Ensure your client is configured to point at the server and that you provide the required API key for WaveSpeed access.

How to install

Prerequisites: Python 3.11 or newer and a WaveSpeed API key.

# Install the MCP server package
pip install wavespeed-mcp

Running the server

Start the WavespeedMCP server in your environment, supplying your WaveSpeed API key. The server will listen for MCP clients to submit tasks and provide progress updates.

wavespeed-mcp --api-key your_api_key_here

Claude Desktop integration (optional)

If you want Claude Desktop to use WavespeedMCP tools, generate a configuration file that links Claude Desktop to the server tools.

python -m wavespeed_mcp --api-key your_api_key_here --config-path /path/to/claude/config

Configuration and runtime behavior

WavespeedMCP can be configured via environment variables, command-line arguments, or a configuration file. You can control the API host, base path for outputs, how resources are returned, logging, and timeouts.

Security and best practices

Keep your WaveSpeed API key secure. Do not share the key in logs or with unauthorized processes. Use separate keys for development and production, and restrict network access to the MCP server as appropriate.

Notes and troubleshooting

If you encounter timeouts, adjust the per-request timeout and total wait-for-result timeout to fit your network and job complexity. Enable console or file logging to diagnose issues, and verify that the server process has the necessary permissions to write outputs to the configured base path.

Configuration details

Environment variables and runtime options control API access, output destinations, and endpoints. Refer to the listed variables to tailor the server to your environment.

Available tools

text_to_image

Generates an image from a text prompt with options for image-to-image, inpainting, and LoRA models.

image_to_image

Applies transformations or enhancements by using an input image as the starting point for generation.

video_generation

Creates dynamic videos from static images with configurable motion parameters and output settings.

resource_handling

Supports multiple resource output modes including URL, local file, and base64 representations.

Built by

VeilStrat

AI signals for GTM teams