Topaz Labs MCP Server (topaz-mcp) is an MCP (Model Context Protocol) server launched by Topaz Labs, used to let LLMs and AI assistants call Topaz Labs' AI image enhancement capabilities through the MCP protocol. Topaz Labs itself provides AI image and video enhancement software for photographers, filmmakers, and developers, and this MCP server connects this capability to AI assistant workflows.
Purpose: To provide image enhancement capabilities for AI assistants such as Claude, enabling them to directly perform AI processing on images during conversations or automated workflows, including upscaling, denoising, and sharpening.
Main capabilities:
- Supports multiple Topaz Labs AI models; those explicitly mentioned in the materials are Standard V2, Wonder 2, Bloom, and Recover 3.
- Uses the Standard V2 model by default and automatically upscales images to 2K.
- Size control options: precise dimensions can be specified via { width, height }; proportional scaling via { width } or { height }; a multiplier via { scale }. If unspecified, images below 2K are automatically upscaled to a 2K width.
- Input methods: a file path or URL must be provided; images uploaded directly in chat cannot be processed.
- Transport: uses stdio transport, supported by all mainstream MCP desktop clients. HTTP streaming transport (for ChatGPT, remote agents, Web/mobile clients) is not yet supported, but may be added in future versions.
Usage:
- Register at developer.topazlabs.com and obtain an API key (paid).
- Add the server to the MCP client configuration. For example, the Claude Desktop configuration file path: on macOS it is ~/Library/Application Support/Claude/claude_desktop_config.json, and on Windows it is %APPDATA%\Claude\claude_desktop_config.json.
- Issue instructions in the conversation, for example: "Enhance ~/Desktop/photo.jpg with Wonder 2", and the AI assistant will call the image-enhance tool to complete the processing.
Applicable scenarios: Suitable for photographers, designers, and developers who need to enhance local images in batches or on demand with the help of an AI assistant. Because it requires file system access, this tool works best in environments with file system access capabilities such as Claude Cowork, and cannot be used in Claude Chat mode (without file system access).
Limitations:
- Only supports images; video enhancement is not yet supported.
- Requires a paid API key.
- Images are sent to the Topaz Labs API for cloud processing.
- Cannot process images uploaded directly in chat; a file path or URL must be provided.
- Does not support HTTP streaming transport, so it cannot be used for ChatGPT, remote agents, Web/mobile clients.
This project uses the MIT license and is maintained by Topaz Labs.
