Interior design and architecture MCP servers connect AI agents to CAD software, BIM platforms, 3D modelers, and design tools. Instead of manually navigating complex CAD interfaces or switching between modeling applications, these servers let you create, modify, and analyze designs through natural language via the Model Context Protocol.
This review covers CAD control, BIM automation, 3D modeling, and architectural design tools — AutoCAD, FreeCAD, Blender, SketchUp, Rhino/Grasshopper, Revit, Fusion 360, and OpenSCAD. For general-purpose 3D file processing, see our 3D Printing review if available.
The headline findings: Blender MCP is the star of the category with 17.6k GitHub stars — one of the most popular MCP servers in existence. Every major CAD platform now has MCP support, from AutoCAD to Rhino to Revit. Rhino/Grasshopper has the most implementations with 5+ independent servers reflecting the computational design community’s enthusiasm. FreeCAD-MCP offers 57 tools including Docker headless execution and AI-powered 3D generation. The glaring gap is actual interior design — room layout, furniture placement, and color palette tools don’t exist yet.
Blender
ahujasid/blender-mcp — The Most Popular MCP Server for 3D
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| blender-mcp | 17,600+ | Python | — | 18+ |
The dominant Blender-to-AI bridge with massive community adoption:
- Two-way communication — connects Claude AI to Blender through a socket-based server
- Object manipulation — create, modify, and delete 3D objects with natural language
- Material control — apply and modify materials and colors
- Scene inspection — get detailed information about the current Blender scene
- Geometry nodes — status query, node info retrieval, node network creation
- Asset libraries — Polyhaven categories/search/download, Sketchfab model search/preview/download
- AI 3D generation — Hyper3D/Hunyuan3D model generation, polling, and import
- Viewport screenshots — capture and share what Blender is rendering
- Code execution — run arbitrary Python code in Blender from Claude
At 17.6k stars, this is one of the most popular MCP servers across all categories, period. The breadth of integrations — from Polyhaven assets to Hyper3D generation — makes it genuinely useful for both hobbyists and professionals.
seehiong/blender-mcp-n8n — Blender via n8n Automation
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| blender-mcp-n8n | — | — | — | 45+ |
Automates Blender 3D modeling from n8n workflows:
- 45+ tools for modeling, materials, lighting, rendering, and animation
- n8n integration — trigger Blender operations from automation workflows
- Comprehensive coverage — goes beyond basic object creation into lighting and rendering
An interesting approach that connects Blender to the broader automation ecosystem rather than just direct AI chat.
AutoCAD
puran-water/autocad-mcp — AutoLISP Execution and P&ID Symbols
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| autocad-mcp | 159 | Python | — | 8 |
The most mature AutoCAD MCP server with industrial features:
- Freehand AutoLISP execution — turns the server from a fixed command set into an extensible automation platform
- 8 consolidated tools — drawing, undo/redo, file operations
- P&ID symbol library — 600+ ISA 5.1-2009 standard symbols for process and instrumentation diagrams
- Dual backends — File IPC (requires AutoCAD) or ezdxf (headless DXF generation without AutoCAD)
- Focus-free dispatch — operates without needing AutoCAD to be the active window
- Robust IPC — ESC prefix and UTF-8 fallback for reliable communication
The ezdxf headless backend is particularly notable — you can generate DXF files without having AutoCAD installed at all.
AnCode666/multiCAD-mcp — One Server, Four CAD Platforms
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| multiCAD-mcp | 2 | Python | — | 46 |
Cross-platform CAD control via AI assistants:
- 46 MCP tools — comprehensive drawing, layer, entity, and file management
- Multi-platform — supports AutoCAD, ZWCAD, GstarCAD, and BricsCAD
- Batch operations (v0.1.1+) — 60-70% fewer API calls when managing multiple items
- Natural language — “Draw a red circle at 50,50 with radius 25” or complex instructions
- Windows only — uses COM technology for CAD communication
The cross-platform approach is smart — supporting 4 CAD platforms through one MCP server means you’re not locked into a single vendor.
daobataotie/CAD-MCP — Natural Language CAD Control
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| CAD-MCP | — | — | — | Multiple |
CAD control via natural language instructions:
- Multi-CAD support — AutoCAD, GstarCAD (GCAD), and ZWCAD
- Natural language processing — combines NLP with CAD automation
- Drawing operations — create and modify drawings through conversational commands
chichicaste/mcp-autocad-server — Learning-Focused AutoCAD Integration
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| mcp-autocad-server | — | — | — | Multiple |
Natural language interaction with AutoCAD:
- Create, modify, and analyze drawings through LLMs like Claude
- CAD element data storage — stores and queries element data
- Learning reference — the author notes this is currently for learning reference, with basic communication implemented but tool functions not yet complete
FreeCAD
proximile/FreeCAD-MCP — The Most Feature-Rich CAD MCP Server
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| FreeCAD-MCP | — | Python | — | 57 |
A powerhouse CAD server with AI generation capabilities:
- 57 MCP tools — the highest tool count of any CAD MCP server
- Docker headless execution — containerized FreeCAD with optional VNC GUI access
- Multi-AI support — works with Claude, GPT-4o, and Gemini
- TRELLIS.2 integration — image-to-3D model generation
- ComfyUI integration — text-to-image generation for design concepts
- Vision AI analysis — Cosmos VLM for model analysis, SAM3 for segmentation
- Gradio web interface — debugging UI with request tracking and image gallery
- Cloudflare tunnels — public URL access without port forwarding
- Dual GPU support — optimized for dual 24GB 3090 setup with automatic quantization
This goes far beyond basic CAD control — it’s essentially an AI-powered design studio with 3D generation, vision analysis, and headless rendering all in one package.
bonninr/freecad_mcp — Direct Claude-FreeCAD Connection
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| freecad_mcp | — | Python | — | Multiple |
Connects FreeCAD to Claude AI through MCP:
- Direct interaction — Claude controls FreeCAD for prompt-assisted CAD 3D design
- MCP-ready — works with Claude and other MCP-compatible tools like Cursor
lucygoodchild/freecad-mcp-server — Cross-Platform FreeCAD Basics
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| freecad-mcp-server | — | Python | — | Multiple |
Basic FreeCAD operations via MCP:
- Geometry creation — basic 3D shapes and objects
- Boolean operations — union, intersection, subtraction
- Custom script execution — run FreeCAD Python scripts
- Cross-platform — Windows, macOS, and Linux
spkane/freecad-addon-robust-mcp-server — MCP Bridge Workbench
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| freecad-addon-robust-mcp-server | — | Python | — | Multiple |
FreeCAD Robust MCP Suite with dedicated workbench:
- MCP Bridge Workbench — integrates directly into FreeCAD’s UI
- Docker support — available as a Docker image for containerized use
- Addon architecture — installs as a FreeCAD addon for tight integration
contextform/freecad-mcp — Open-Source Automation
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| freecad-mcp | — | — | — | Multiple |
Open-source MCP server for FreeCAD automation. Another option in the growing FreeCAD MCP ecosystem.
SketchUp
mhyrr/sketchup-mcp — Architectural Modeling and Woodworking
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| sketchup-mcp | 182 | Ruby/Python | — | Multiple |
SketchUp-to-AI bridge with unique woodworking tools:
- TCP socket connection — two-way communication between Claude and SketchUp
- Component manipulation — create, modify, delete, and transform components
- Material control — apply and modify materials
- Scene inspection — detailed information about the current scene
- Selection handling — work with selected objects
- Ruby code evaluation — execute arbitrary Ruby code directly in SketchUp
- Woodworking joints — mortise and tenon, dovetail, finger joint (unique to this server)
- Planned features — lap joints, miter joints, dowel joints, pocket holes, grain direction, wood species libraries
The woodworking joinery tools make this uniquely valuable for furniture design and crafting — no other CAD MCP server offers this level of woodworking-specific functionality.
BearNetwork-BRNKC/SketchUp-MCP — Traditional Chinese Localized
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| SketchUp-MCP | — | — | — | Multiple |
SketchUp MCP integration with Traditional Chinese localization. A community fork targeting Chinese-speaking users.
Rhino & Grasshopper
jingcheng-chen/rhinomcp — The Leading Rhino MCP Server
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| rhinomcp | 274 | Python | — | Multiple |
The most popular Rhino-to-AI bridge:
- Object manipulation — create, modify, and delete 3D objects
- Document inspection — detailed information about the Rhino document
- Script execution — run Python scripts directly in Rhino
- Object selection — filter by name, color, category with logic operators
- Layer management — get, set, create, or delete layers
- JSON-based TCP protocol — clean, simple communication protocol
- Available on PyPI — easy installation via pip
At 274 stars, this leads the Rhino MCP ecosystem with a clean, well-documented implementation.
veoery/GH_mcp_server — Direct Rhino + Grasshopper Interaction
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| GH_mcp_server | — | Python | — | Multiple |
LLM-to-Rhino/Grasshopper bridge for designers:
- 3dm file analysis — analyze existing Rhino files
- 3D modeling — direct modeling in Rhino
- GHPython generation — automatically generate Grasshopper Python scripts based on user guidance
- Designer-focused — built for design workflows, not just technical CAD
dongwoosuk/rhino-grasshopper-mcp — ML Auto-Layout
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| rhino-grasshopper-mcp | — | Python | — | Multiple |
AI-powered Rhino/Grasshopper with machine learning layout:
- ML-based automatic layout optimization — uses machine learning for spatial arrangement
- Direct AI integration — brings AI capabilities into Rhino/Grasshopper workflows
alfredatnycu/grasshopper-mcp — Component Creation and Intent Recognition
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| grasshopper-mcp | — | — | — | Multiple |
Bridging server between Grasshopper and Claude Desktop:
- Component creation — create and connect Grasshopper components via MCP
- High-level intent recognition — understands design intent, not just commands
- MCP standard — follows the protocol specification for reliable integration
reer-ide/rhino_mcp — Multi-Tool Rhino Integration
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| rhino_mcp | — | — | — | Multiple |
Connects Rhino, Grasshopper, and more to Claude AI:
- Prompt-assisted 3D modeling — create and manipulate scenes through conversation
- Multi-tool — integrates Rhino, Grasshopper, and additional tools
Revit / BIM
revit-mcp/revit-mcp — AI-Powered Building Information Modeling
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| revit-mcp | 18 | C#/Python | — | Multiple |
The only major BIM platform with MCP support:
- CRUD operations — create, read, update, delete Revit elements through AI
- Plugin architecture — revit-mcp-plugin receives AI commands within Revit
- Command sets — revit-mcp-commandset wraps Revit external events as executable commands
- Extensible — developers can add functionality to the command set or create custom work sets
- MCP protocol clients — works with Claude, Cline, and other MCP-compatible clients
Revit is the industry standard for BIM in architecture, and having MCP support — even early-stage — opens up AI-assisted building design workflows that were previously impossible.
Fusion 360
AuraFriday/Fusion-360-MCP-Server — Thread-Safe Fusion Control
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| Fusion-360-MCP-Server | — | Python | — | Multiple |
Thread-safe Fusion 360 control via MCP:
- Thread-safe architecture — all API calls happen on main thread to prevent crashes
- Multi-AI support — works with Claude, ChatGPT, and other MCP clients
- Parametric modeling — leverages Fusion 360’s parametric design capabilities
mycelia1/fusion360-mcp-server — Script Generation and Execution
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| fusion360-mcp-server | — | Python | — | Multiple |
Generate and execute Fusion 360 scripts via AI:
- Script generation — AI creates Fusion 360 scripts from natural language
- Direct execution — runs generated scripts in Fusion 360
- Multi-client — works with Claude, Cursor, or any MCP client
Joe-Spencer/fusion-mcp-server — Autodesk Resources and Tools
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| fusion-mcp-server | — | — | — | Multiple |
Provides Autodesk resources and tools to AI clients:
- ADSK resources — exposes Autodesk documentation and capabilities
- Tool integration — bridges Fusion 360’s API with MCP protocol
OpenSCAD
quellant/openscad-mcp — Multi-Perspective Rendering
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| openscad-mcp | — | Python | — | 3 |
Production-ready OpenSCAD rendering via MCP:
- Single view rendering — render individual perspectives with customizable camera
- Multi-perspective rendering — front, back, left, right, top, bottom, isometric views
- Animation support — turntable animations for model visualization
- 300 tests — 80%+ code coverage, 100% integration test success
- Base64 PNG output — renders returned as encoded images
- Async processing — built on FastMCP for non-blocking operation
fboldo/openscad-mcp-server — STL and PNG Output
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| openscad-mcp-server | — | Python | — | Multiple |
Renders PNG previews and STL geometry from OpenSCAD code. Focused on generating both visual previews and printable 3D files.
Specialized Tools
Svetlana-DAO-LLC/cad-agent — Self-Contained Rendering for 3D Printing
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| cad-agent | — | Python | — | Multiple |
AI-driven CAD modeling for 3D printing:
- build123d — uses the build123d Python library for parametric modeling
- Self-contained rendering — the container does all rendering work
- Visual feedback — returns images the agent can interpret
- 3D printing focused — designed for printable model generation
Bigchx/mcp_3d_relief — Image to 3D Relief Models
| Server | Stars | Language | License | Tools |
|---|---|---|---|---|
| mcp_3d_relief | — | Python | — | Multiple |
Converts 2D images into 3D relief models:
- Image-to-STL — transforms flat images into 3D relief surfaces
- 3D printing ready — outputs STL format suitable for printing or rendering
What’s Missing
The biggest gap in this category is interior design itself. Despite the strong CAD/3D modeling coverage, there are no MCP servers for:
- Room layout and space planning — no automated furniture placement, no traffic flow analysis
- Floor plan generation — no text-to-floor-plan tools
- Color palette and material selection — no integration with paint brands, fabric libraries, or finish catalogs
- Furniture catalogs — no IKEA, Wayfair, Pottery Barn, or other retailer APIs
- AR/VR visualization — no augmented reality room preview tools
- Rendering engines — no V-Ray, Lumion, Enscape, or KeyShot MCP servers
- Landscape architecture — no garden design, hardscape planning, or plant databases
- Building code compliance — no zoning, ADA, or fire code checking
- Lighting design — no illumination calculations or fixture selection
- Acoustics — no room acoustic analysis or soundproofing recommendations
- Structural engineering — no load calculations or structural analysis
- Cost estimation — no material quantity takeoffs or budget planning
- SolidWorks — the most widely used professional mechanical CAD has no established MCP server
- Onshape — the cloud-native CAD platform lacks MCP integration
Bottom Line
The interior design and architecture MCP server category is exceptionally strong on the CAD/3D modeling side but almost completely empty on the design decision-making side. If you need AI-controlled CAD — Blender, AutoCAD, FreeCAD, SketchUp, Rhino, Revit, or Fusion 360 — you have excellent options with active development and significant community adoption. Blender MCP at 17.6k stars is among the most popular MCP servers in existence, and the Rhino/Grasshopper ecosystem has 5+ independent implementations showing genuine demand from the computational design community.
But if you’re looking for AI-assisted interior design — “help me arrange furniture in my living room” or “suggest a color palette for this bedroom” — MCP servers can’t help you yet. This represents a significant opportunity: the interior design market is massive, and the gap between powerful CAD automation tools and accessible design tools is exactly where new MCP servers could add the most value.
Rating: 4/5 — Outstanding CAD/3D tool coverage with strong community adoption. The Blender, Rhino, and AutoCAD ecosystems alone justify a high rating. Points deducted for the complete absence of interior design workflow tools, rendering engine integration, and building code compliance checking.
This review was last edited on 2026-03-16 using Claude Opus 4.6 (Anthropic).