Embody, developed by Dylanroscover, is a free mcp program built with Python that integrates AI assistants directly into the TouchDesigner environment. It functions by creating a per-project Python environment aligned with TouchDesigner's interpreter. Distributed as a .tox file, Embody's core component, Envoy, serves as an embedded Model Context Protocol server. This allows AI clients to interact with live TouchDesigner sessions, enabling advanced automation and creative workflows for artists and developers.

What does Embody connect to?

Embody operates within TouchDesigner, a node-based visual development platform. Its central component, Envoy, functions as an embedded Model Context Protocol (MCP) server. The program also includes specific tools designed for externalizing TouchDesigner networks into text files. Additionally, it can convert these networks into a human-readable YAML format known as TDN, which aids in understanding and managing project structures outside the visual interface.

Core features and tools

Through its embedded Envoy MCP server, Embody exposes callable tools for AI assistants to interact with a live TouchDesigner session. These tools enable actions such as creating operators, wiring connections, setting parameters, writing extensions, reading cook errors, running Python code, and capturing rendered output. The program offers a substantial collection of these tools, with documentation mentioning either 54 MCP tools (plus 4 bridge meta-tools) or 49 tools for various operations.

Who is it for?

The primary audience for Embody consists of TouchDesigner users, including artists, developers, and creators. These individuals seek to employ AI assistants to accelerate their development processes and enhance creative workflows. It is designed for those who want to automate network construction, quickly iterate on designs, and manage project versions more effectively. Skip this if you require network access beyond local-only connections.

How to get started

No credentials are required. The server binds to HTTP for local-only access.

Use cases

• AI-driven network construction, where users describe desired networks in natural language for automated creation in TouchDesigner.
• Rapid iterative design and modification, allowing quick adjustments or new approaches based on AI suggestions.
• Version control and state management, by externalizing TouchDesigner networks into human-readable TDN files for use with standard version control systems.

Embody offers a powerful integration for TouchDesigner users, enabling AI assistants to directly manipulate and manage projects. Its local-only HTTP server ensures secure, contained operation, while its extensive toolset supports complex AI-driven workflows. The ability to externalize networks to TDN files significantly improves version control. This program is a valuable addition for creators seeking to accelerate their development with AI. Start by describing a desired network in natural language to an AI assistant.