> For the complete documentation index, see [llms.txt](https://docs.m-xr.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.m-xr.com/marso-studio/mcp/what-marso-mcp-does.md).

# What Marso MCP does

Marso MCP connects an AI agent to Marso Studio. Add it to Claude, ChatGPT, Codex or Claude Code, and you can ask the agent for 3D assets in a chat. It builds them on Marso Studio's node graph, in your own account.

### What your agent can do

* Generate images of single objects or whole scenes, and edit images with a prompt.
* Detect the objects in a scene image, and let you choose which ones to keep.
* Turn images into 3D models, and give them predicted PBR materials.
* Give predicted PBR materials to 3D models you upload.
* Change the textures of a 3D model with a prompt.
* Work across many objects at once, such as every object in a scene.
* Open and add to your projects, browse your Library, show results in the chat, and give you download links.
* Work in your open Blender or Isaac Sim through the Marso add-ons. See [Blender and Isaac Sim](/marso-studio/mcp/blender-and-isaac-sim.md), and [Blender with your AI agent](/marso-studio/plugins/m-xr-plugins/blender-with-your-ai-agent.md) to have your agent model in Blender too.

### Where your work goes

Your agent works in your Marso Studio projects, the same ones you open at [marso.app](https://marso.app/). It uses the project you're working on, or starts a new one named after what you asked for. Open the project any time to check what the agent did and carry on by hand. What it makes on a project shows in your Library too: see [Library & collections](/marso-studio/reference/library-and-collections.md).

Your agent can add nodes to a project and connect them, but it can't delete a project or a node.

### What you need

* A Marso Studio account and an API key. See [Connect your agent](/marso-studio/mcp/connect-your-agent.md).
* An agent that can use MCP connectors: Claude, ChatGPT, Codex or Claude Code.
* A paid plan or the trial. Marso MCP works through the Marso Studio API, which the Free plan doesn't include, and a key made during the trial stops working when the trial ends. See [Plans](/marso-studio/reference/credits-and-billing.md#plans).

### Credits

Marso MCP spends credits from your Marso Studio balance, at the same prices as the app. See [Credits & billing](/marso-studio/reference/credits-and-billing.md).

Before a larger step, your agent checks the price against your balance, and by default it asks you first. You can tell it how much it may spend without asking ("go ahead, up to 500 credits"), give it a standing allowance, or tell it to run straight through. These are instructions your agent follows. Marso charges each run what it costs.

### Your own files

Marso MCP runs online, so it can't open files on your computer. To use your own image or 3D model, your agent gives you an upload link. Drop the file on that page, then go back to the chat: the agent carries on by itself. Claude Code can upload files from your computer for you. See [Marso MCP troubleshooting](/marso-studio/mcp/marso-mcp-troubleshooting.md) for file types and sizes.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.m-xr.com/marso-studio/mcp/what-marso-mcp-does.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
