English | 中文

Website · Marketplace · Showcase · For Agents · Ecosystem

Skill-driven infrastructure for agents to operate real creative tools.

DCC-MCP (Digital Content Creation Model Context Protocol) does not build or prescribe an agent. It connects agents to a growing ecosystem of desktop DCCs, game engines, 2D tools, production systems, asset providers, profilers, and custom studio hosts through shared discovery, execution, safety, and operations contracts.

Install the public Skill that matches the task. Operating an existing DCC only needs dcc-mcp; the creator Skills are focused authoring guides.

openclaw skills install @loonghao/dcc-mcp
openclaw skills install @loonghao/dcc-mcp-creator
openclaw skills install @loonghao/dcc-mcp-skills-creator
# Direct ClawHub CLI for other Agent Skills hosts:
npx --yes clawhub@0.23.1 install @loonghao/dcc-mcp
npx --yes clawhub@0.23.1 install @loonghao/dcc-mcp-creator
npx --yes clawhub@0.23.1 install @loonghao/dcc-mcp-skills-creator

Start a new agent turn after installation. The default workflow is dcc-mcp + dcc-mcp-cli: inventory with dcc-mcp-cli list, inspect startup state with dcc-mcp-cli doctor, then search -> describe -> call. The Skill also teaches the agent to preserve trace evidence, analyze the failing layer, and route a sanitized bug report to the owning Skill, adapter, or Core project. Use the dcc-mcp-cli verified installation guide when the CLI is not already available.

The shortest DCC agent demo asks a model to write and run a mayapy, hython, or Blender Python script. A production pipeline cannot depend on getting the right script from the model on every turn.

Repeated code generation costs tokens and makes results vary with the model, prompt, and context. Adapter authors also end up rebuilding the same transport, main-thread dispatch, validation, process lifecycle, instance routing, and diagnostics for every host.

DCC-MCP moves that common engineering into one reusable framework:

Agents and models will change. Studio interfaces, permission boundaries, and pipeline knowledge still need to be maintained. The framework turns those investments into reusable engineering assets.

We reuse MCP as an industry-standard agent interface instead of inventing a private protocol. The framework is not limited to MCP: a stable Python, C++, HTTP, command-port, or native plugin interface can be integrated under the same control plane.

We also include useful vendor capabilities instead of replacing them. Unreal Engine 5.8 introduced an experimental official MCP server and Toolset Registry. Our unreal-official-mcp Skill enables and bridges those official toolsets without redistributing Epic's plugin or changing its tool names and schemas. Vendor tools, DCC-MCP tools, and studio tools can share one agent-facing workflow.

Adding an AI-facing interface to a new tool is usually straightforward. The harder problem is making years of existing tools usable by agents when they have no API, cannot be modified, or expose part of a workflow only through a window or modal dialog.

dcc-mcp-core provides a scoped Computer Use-style capability called UI Control. It uses a deterministic snapshot -> find -> act -> wait -> verify loop over Qt, native accessibility, webviews, or host-specific UI backends. Native Skills/APIs remain preferred; whole-desktop access is denied by default, and UI actions stay scoped, policy-checked, auditable, and verifiable. See the UI Control workflow guide.

A Skill is not just a script in another folder. It packages proven pipeline knowledge as a versioned, typed, testable, and distributable operation.

A lower-cost model may struggle to invent scene-editing logic from scratch but can reliably select a well-described tool and provide validated arguments. Skills reduce repeated code generation, token use, and model-dependent variance. Large studios can distribute different Skill sets by project, department, and production stage without rebuilding an adapter.

For TDs and TAs, Skills are the shortest path from an internal requirement to a reusable capability. Keep host connectivity, main-thread execution, routing, safety, and observability in Core and the adapter; package project naming, scene checks, asset preparation, publish gates, cache/export rules, and review hand-offs in SKILL.md, tools.yaml, and existing studio scripts. The result can be tested, project-scoped, and distributed through a public or private Marketplace without forking the control plane.

The official Marketplace makes those optional capabilities searchable, installable, upgradeable, and suitable for private studio catalogs too. Browse real outputs and reusable prompts, or inspect the catalog source.

The official catalog currently publishes all of the following optional capabilities. Agents should search the live catalog before installation:

dcc-mcp-cli marketplace search --query "<capability>" --limit 20
dcc-mcp-cli marketplace install <package_name> --dcc <dcc_name>

The Gateway Admin UI exposes calls, traces, logs, health, statistics, and usage activity. Teams can see which tools agents selected, where calls failed, and which Skills are used most often. That evidence can drive a practical feedback loop: improve a description, schema, or implementation, then verify the result against real calls.

Browse the DCC-MCP repositories for more adapters, asset providers, UI automation, profilers, and production Skills.

The boundary is deliberate: the core does not replace host-specific pipeline semantics or promise safe rollback where a DCC exposes no transaction API. Those responsibilities stay in adapters and pipeline Skills.

If you want DCC agents to move from demos to reusable, deployable, and observable production tools, try the project and give it a Star.

Issues and pull requests are welcome. Please report bugs, integration gaps, and production workflow needs in the relevant repository. Focused adapters, Skills, tests, documentation, and interoperability improvements are especially useful.