RhinoAI vs. RhinoMCP: Two ways to use AI with Rhino
Rhino can be connected to AI tools in two different ways. Both options give an AI assistant access to core Rhino modeling tools (commands, scripting, and Grasshopper), while RhinoMCP adds external process control. The main difference is where the AI runs, and how interactive it is with modeling.
RhinoAI panel (inside Rhino)
Run the AI command (formerly AIPanel) to open a chat panel right inside Rhino. The AI connects directly to Rhino, Grasshopper, and the Script Editor.
- Works with you on the model, using what you’ve selected and what’s on screen
- Runs scripts and commands right where you’re modeling
- Can use other MCP servers (add-on AI tools) already set up on your computer
- Can turn a prompt into a macro or a new command
- Nothing extra to install or set up
The bottom line: RhinoAI is your hands-on modeling partner. With zero setup required, it keeps the AI right beside your cursor so you can model, iterate, and experiment together in real time without ever leaving your viewport.
RhinoMCP server (outside Rhino)
Run the MCPConnect command in Rhino to start the connection, turning Rhino into an MCP tool server. From there, you work with an AI in whichever non-Rhino “surface” fits your setup:
| Surface | MCP-Compatible Clients |
|---|---|
| Terminal | Anthropic Claude Code CLI (claude), OpenAI Codex CLI (codex), Google Antigravity CLI (agy), xAI Grok Build CLI (grok), Block Goose CLI (goose), Cognition Devin CLI (devin) |
| Desktop | Claude Desktop, ChatGPT Desktop, Antigravity App, Goose Desktop, LM Studio |
| IDE | VS Code, Antigravity IDE, Cursor, Devin Desktop (formerly Windsurf), Zed |
- Can start, close and restart Rhino for you
- Can work with several copies of Rhino 8 and Rhino 9 at the same time
- Can open, save and switch between files
- Can load, test and debug plugins, and keep working even if Rhino crashes
- Can combine Rhino with the app’s other tools and connections in one conversation
- Can turn a prompt into a macro or a new command
The trade-off: RhinoMCP requires a separate AI client and setup, but moving the AI outside the Rhino window turns it from an in-canvas assistant into an autonomous director: it keeps your modeling viewport completely clean, stays alive even if Rhino crashes, and can coordinate Rhino alongside the rest of your computer’s tools.