RhinoAI and RhinoMCP are two distinct tools for working with AI in Rhino 3D. RhinoAI puts a conversational assistant inside Rhino, alongside your modelling work. RhinoMCP connects an external AI application to Rhino so that it can direct modelling tasks, scripts and wider workflows.

They share underlying technology, but serve different ways of working. RhinoAI is the in-Rhino modelling partner; RhinoMCP is the connection for an assistant working outside Rhino. This guide brings together the key differences, getting started and practical troubleshooting.

RhinoAI: An Assistant Inside Rhino

RhinoAI provides a chat panel within Rhino, with access to Rhino, Grasshopper and the Script Editor. You can work on the model yourself while asking the assistant to help with selected objects, run scripts and commands, or develop a repeatable task into a macro or command.

This suits hands-on design work: exploring a shape, refining a script or asking for help with a modelling operation while keeping your attention on the viewport.

McNeel's current announcement uses the AI command to open the panel; some builds and documentation use AIPanel. Choose an available agent in the panel and complete its configuration and sign-in. Follow the instructions for your installed build, as commands and supported agents evolve.

RhinoMCP: Connect an External AI Assistant

RhinoMCP lets an AI assistant in a separate desktop application, terminal or development environment use Rhino as a tool. MCP stands for Model Context Protocol, the connection through which the assistant can request operations rather than simply give you text instructions to follow.

Depending on the supported client and configuration, an external assistant can create geometry, run scripts, work with Grasshopper, open and save documents, and help test or debug plugins. The external workflow can also coordinate Rhino with the assistant's other tools.

This suits automation and development work, or anyone who already works in an external AI application. You direct the assistant there, rather than chatting inside Rhino's modelling window.

Which Should You Start With?

Start with RhinoAI if you want an assistant beside you while you model. Choose RhinoMCP if your workflow centres on an external AI application, scripting or coordinating tasks across tools. The choice is about where you work and how you want to interact, rather than one being a replacement name for the other.

Getting Started

For RhinoAI, use the AI panel provided by your supported Rhino setup and configure an available agent. McNeel's AI panel instructions explain the panel workflow.

For the external RhinoMCP connection, McNeel's current command documentation describes MCPConnect: install your chosen AI application, run the command in Rhino and use the connection dialogue to add its configuration. Then follow the application's instructions to load the updated configuration.

Separate package-based installation guidance also exists for the Rhino MCP Platform. Use the official Rhino MCP documentation for the installation route that matches your Rhino build and chosen client. Panel configuration and external-client installation are separate workflows.

A Useful First Task

Try a small, inspectable task on a copy of a model. For example: “Create three rectangular panels on a new layer called Panel Study. Make each panel 1,000 mm wide and 2,000 mm high, with a 100 mm gap. Keep the originals unchanged.”

Give the assistant dimensions, units, object types and a clear scope. For editable surface modelling, specify the construction you want, such as lofting curves or revolving a profile. Inspect the resulting geometry and dimensions before extending the task.

Grasshopper, Plugin Data and Image References

AI-assisted scripting can help with repetitive modelling and parametric work. Tool availability depends on your installed setup, so start with a small definition and review its inputs, connections and output.

For specialist plugins, distinguish access to Rhino geometry from access to the plugin's own data and commands. Schedule extraction, clash checks and automated dimensioning are workflows to develop and validate against a clear brief, rather than universal ready-made features of either product.

If your chosen assistant accepts images, a reference photograph can help communicate proportions and design intent. Supply known dimensions and additional views wherever possible, and treat generated geometry as a starting point to measure and refine.

Using AI in Professional Work

Keep tasks scoped, save a copy before significant changes and review geometry, dimensions and scripts before production use. For client projects, review your AI provider's data-handling terms and your organisation's requirements before sharing project information.

Choose your Rhino licence for the way you and your team work, and review the separate terms of your AI provider. Our Rhino 3D licensing guide explains the main licence-management options.

For the latest distinction between the two tools and their supported workflows, see McNeel's RhinoAI and RhinoMCP announcement.