Ashfox is a free Model Context Protocol (MCP) server tool developed by Sigee Min, designed for AI-native low-poly 3D asset generation. Built primarily with TypeScript and JavaScript, it functions as an asset compiler for various stages of 3D content creation, including modeling, texturing, rigging, and animation. Users interact with a browser-based workbench to create game-ready exports. Ashfox aims to simplify 3D asset development by integrating AI agents into the design workflow, providing a new approach to digital content production for game developers and artists.

What does Ashfox connect to?

Ashfox connects to an AI agent to compile "Intent Programs" into canonical low-poly assets. This core connection enables its function as an asset compiler for 3D modeling, pixel texturing, rigging, animation, and game-ready exports. As an MCP server, Ashfox adheres to standard transport mechanisms like stdio and Streamable HTTP. Given its browser-based workbench, Streamable HTTP is the probable transport method, facilitating interaction with a browser-capable agent by fetching an "agent-manifest.json" via direct HTTP requests.

Core features and tools

Ashfox's primary capability is AI-driven asset generation and compilation. Users interact by describing assets in ordinary language to an AI agent, observing the build in a viewport, and requesting changes in natural language. It offers various output formats, referred to as export adapters, for delivery artifacts. These include Java block, GeckoLib 5, Bedrock, GLB, or glTF. Ashfox does not expose discrete, named callable tools in the traditional MCP sense; its interaction model centers on the conversational AI asset creation process itself.

Who is it for?

Ashfox is for users involved in 3D asset creation, particularly those developing for games, who wish to utilize AI for modeling, texturing, rigging, and animation. It suits creators looking for an AI-native workbench to generate low-poly assets efficiently. This tool is ideal for designers seeking to describe assets conversationally and iteratively refine them. Skip this if you require a traditional MCP server with a documented list of discrete, named callable tools for direct programmatic invocation.

How to get started

No credentials are required to begin using Ashfox. Its design as a local-first, browser-based workbench means there is no explicit authentication model or account creation needed for its primary use.

Use cases

• AI-driven 3D asset generation: Describe a desired asset in natural language to an AI agent, then iteratively refine its design and properties through ongoing conversational interaction.
• Exporting finished assets: Deliver game-ready artifacts by compiling them into various specialized formats, such as Java block, GeckoLib 5, Bedrock, GLB, or glTF.

Ashfox provides a distinct approach to 3D asset creation by integrating AI-driven compilation directly into a browser-based workbench. Developed by Sigee Min, this free MCP server simplifies the generation of low-poly models, textures, rigging, and animations through natural language interaction. It supports multiple game-ready export formats, making it valuable for game developers and 3D artists.