dcc-mcp-core

What it is and what it does

dcc-mcp-core is a foundational library that bridges agents and AI models to desktop creative applications—DCCs like Maya, Blender, and Houdini, as well as game engines and 2D tools—through a unified MCP (Model Context Protocol) and REST interface. It handles the infrastructure that would otherwise be rebuilt for each adapter: main-thread dispatch, multi-instance routing, tool discovery, structured result validation, and lifecycle management.

The package centers on Skills—versioned, testable, distributable operations that encode pipeline knowledge without requiring repeated code generation. A studio TD or TA can compose Skills from SKILL.md, tools.yaml, and existing scripts, scoped to a project or team, then distribute them through a marketplace. The control plane keeps ownership of host connectivity, safety, and observability while letting teams customize the actual workflow logic. It also provides a bounded Computer Use-style UI Control capability for legacy tools that have no programmatic API.

Use it for:

• Control a running Maya, Blender, or Houdini session from an agent or CI job using the CLI and REST gateway.
• Expose a custom DCC adapter or legacy tool over MCP/REST without rebuilding transport and dispatch infrastructure.
• Package and distribute reusable Skills (e.g., asset providers, rigging, procedural authoring) through a marketplace.
• Route agent requests to the correct DCC instance and handle main-thread affinity, async jobs, and cancellation.
• Integrate UI automation and Computer Use workflows for tools that lack a clean programmatic API.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

dcc-mcp-core provides a Rust-powered control plane that connects agents to desktop DCCs, game engines, and creative tools through discoverable MCP and REST capabilities, handling main-thread dispatch, routing, and structured tool execution.

Yes, if you are building or extending an agent-driven creative pipeline. dcc-mcp-core is actively maintained (released 1 day ago), MIT-licensed, and eliminates the need to rebuild common DCC integration infrastructure. Beta status and a growing ecosystem of adapters make it a solid foundation for studios and tool developers. Install friction is moderate due to compiled wheels, but Python 3.7+ support is broad. No known security vulnerabilities.

Install

dcc-mcp-core on PyPI

pip install dcc-mcp-core

uv add dcc-mcp-core

poetry add dcc-mcp-core

Installing dcc-mcp-core

Before you install

Released 1 day ago with active maintenance. Medium install friction due to compiled wheels (cp37–cp37m, win_amd64 variants present), but supports Python 3.7–3.14 and depends on only one runtime package (dcc-mcp-server). Beta status and recent release suggest ongoing development.

License in practice

MIT license (permissive) places no restrictions on commercial or proprietary use, modification, or distribution—standard permissive terms suitable for production integration.

Quickstart

pip install dcc-mcp-core

from dcc_mcp_core import create_skill_server

server = create_skill_server()
server.run()

Requires dcc-mcp-server as a runtime dependency; compiled wheels available for Python 3.7+ on Linux and Windows x86_64.

Verify before relying

• Whether dcc-mcp-server is packaged separately or must be installed alongside dcc-mcp-core.
• Specific DCC versions and platforms supported by each adapter (Maya, Blender, Houdini, etc.).
• Performance characteristics and concurrency limits for multi-instance DCC routing.
• Whether the Admin UI and marketplace integration are included in this package or require separate installation.

Package facts

Evidence: dcc_mcp_core-0.20.6-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; dcc_mcp_core-0.20.6-cp37-cp37m-win_amd64.whl

Keywords: dcc, mcp, maya, blender, houdini, rust, pyo3

Tags

More Libraries packages

Further reading