OpenClaw is a self-hosted, open-source gateway that connects messaging apps like WhatsApp, Telegram, Discord, and iMessage to AI agents. It lets you chat with AI assistants from anywhere. Skills extend what the agent can do, including generating 3D models via the 3D AI Studio skill.

How do I generate 3D models with OpenClaw?

Install the 3D AI Studio skill from ClawHub, set your API key, and start chatting. Just describe what you want in natural language, like 'make me a 3D model of a medieval sword', and the agent generates a GLB file in your workspace automatically. No commands or coding needed.

Is the 3D AI Studio OpenClaw skill free?

The skill itself is free and open source (MIT license). 3D model generation uses credits on 3D AI Studio, starting at 15 credits for TRELLIS.2 and 35 credits for Hunyuan Rapid. New users receive free credits to get started. Failed generations are automatically refunded.

What AI models are available through the skill?

Three AI models are available: TRELLIS.2 by Microsoft for fast, affordable image-to-3D (15-55 credits), Hunyuan Rapid by Tencent for balanced text or image to 3D (35-55 credits), and Hunyuan Pro for highest quality with PBR textures and multi-view support (60-100 credits).

Can I 3D print models generated through OpenClaw?

Yes. The agent can convert your GLB output to STL format automatically. The STL files are ready to load directly into any slicer like Creality Print, Cura, PrusaSlicer, or BambuStudio. Models generated with Geometry mode produce clean hard-surface topology ideal for printing.

How long does 3D model generation take?

TRELLIS.2 generates models in 25 seconds to 4 minutes depending on resolution. Hunyuan Rapid takes 1 to 2 minutes. Hunyuan Pro takes 3 to 6 minutes for the highest quality output. The agent polls automatically and delivers the result when ready.

What file formats does the skill output?

The primary output format is GLB with optional PBR textures. The agent can convert to STL for 3D printing, OBJ for legacy software, FBX for game engines, and other formats. Conversion happens either locally via Blender or through the 3D AI Studio conversion API.

Do I need a GPU or powerful computer?

No. All 3D generation runs in the cloud on 3D AI Studio servers. OpenClaw itself is lightweight and runs on any machine with Node.js. The skill is a Python script using only standard library modules. You only need an internet connection.

How do I install the 3D AI Studio skill?

Run 'clawhub install whale-professor/3daistudio-integration' from your terminal, or manually download the skill folder into your OpenClaw workspace skills directory. Then set the THREE_D_AI_STUDIO_API_KEY environment variable with your API key from 3D AI Studio.

Can I use generated models commercially?

Yes. All 3D models generated through 3D AI Studio are yours to use commercially in games, apps, products, marketing materials, 3D printing, and any other purpose without restrictions.

What is the difference between TRELLIS.2, Hunyuan Rapid, and Hunyuan Pro?

TRELLIS.2 is fastest and cheapest but only accepts images (no text prompts). Hunyuan Rapid supports both text and image input at a balanced cost. Hunyuan Pro offers the highest quality with PBR textures, multi-view input, face count control, and generation modes like LowPoly, Geometry, and Sketch.

Can I generate 3D models from text descriptions?

Yes, using Hunyuan Rapid or Hunyuan Pro. Describe any object in natural language and the model generates a complete 3D asset. TRELLIS.2 is image-only. For best results, include details about materials, style, and intended use in your description.

NewJust released

OpenClaw Integration

Generate 3D Models from Chat with OpenClaw

Describe what you want in natural language and get a production-ready 3D model delivered to your workspace. Text-to-3D, image-to-3D, and direct STL export for 3D printing. All from a chat message.

Free & open source skillFrom 15 credits/modelNo GPU required

View full API documentation

OpenClaw Chat

“Generate a dark fantasy orc warrior bust with tusks, pointed ears, leather armor with metal spikes. Use Hunyuan Pro with PBR textures.”

What is OpenClaw?

OpenClaw is a self-hosted, open-source gateway that connects your messaging apps to AI coding agents. It turns WhatsApp, Telegram, Discord, iMessage, or any browser into a direct interface to powerful AI tools.

Skills are modular packages that extend what the agent can do. The 3D AI Studio skill teaches the agent how to generate 3D models using our API - so you can create production-ready 3D assets just by chatting.

Self-hosted - your data stays on your machine

Works from any messaging app or browser

3,200+ community skills on ClawHub

MIT licensed, free and open source

No GPU or powerful hardware required

OpenClaw generating a 3D model from a chat message

Chat Interface

Generate 3D models in natural language from any messaging app.

Self-Hosted

Runs on your machine. No data leaves your control.

Skill Ecosystem

3,200+ skills on ClawHub. Install the 3D AI Studio skill in seconds.

Open Source

MIT licensed. Read the code, modify it, contribute back.

How to generate 3D models with OpenClaw

From installation to your first 3D model in four steps

01

Install OpenClaw

Install OpenClaw via npm and run the setup wizard. Takes about 5 minutes on any machine.

02

Add the 3D AI Studio Skill

Install the community skill from ClawHub and set your API key. One command, done.

03

Chat to Generate

Open Telegram, Discord, or the web UI. Describe what you want in plain English.

04

Download & Use

The agent generates, polls, and saves your GLB file. Convert to STL, FBX, or OBJ as needed.

See it in action

A real conversation with OpenClaw generating a 3D model from a chat message

OpenClaw Chat

Direct gateway session

connected

Generate a 3D model of a geometric vase with low-poly facets, clean hard-surface topology, good for 3D printing

Generating with Hunyuan Pro 3.1 in Geometry mode...

Model: Hunyuan Pro | Type: Geometry | Est. credits: 60 Submitting generation request... Task ID: f7a2c... [45s] Status: PROCESSING (32%) [120s] Status: FINISHED (100%)

Done! Generated in ~2 minutes.

geometric_vase.glb - 8.6 MB geometric_vase.stl - 23.8 MB (print-ready)

Used Geometry generation type which optimizes for clean hard-surface topology - ideal for architectural and geometric shapes. Both files are in your workspace ready to slice!

Type a message...

Generated via OpenClaw + 3D AI Studio

Real outputs from text and image prompts sent through chat. No post-processing.

PBR on PBR off

Albedo

“anime style female engineer in orange and gray power armor”

PBR

“retro sci-fi ray gun, turquoise red and gold, stylized 3D weapon”

Albedo

“stylized treasure chest with gold coins, padlock, vines and moss”

Output - cartoon skeleton juggler with top hat, bow tie and colorful bones

Albedo

Input - cartoon skeleton juggler with top hat, bow tie and colorful bones

Output - dark fantasy horned orc bust with tusks, beard and glowing eyes

PBR

Input - dark fantasy horned orc bust with tusks, beard and glowing eyes

Output - stylized fantasy rogue girl in hooded red outfit with dual daggers

Albedo

Input - stylized fantasy rogue girl in hooded red outfit with dual daggers

Text to 3D - just describe what you want

Type a description in any connected chat. The agent picks the right model, generates, and delivers your 3D file.

Character Generation

OpenClaw Chat

Direct gateway session

connected

Make a 3D model of a cute yellow electric mouse character with a zigzag lightning bolt tail. Round body, big oval eyes, rosy cheek patches, standing in A-pose. Manifold mesh for 3D printing, ~100mm tall with a flat base.

Using Hunyuan Pro 3.1 for best character detail...

Model: Hunyuan Pro 3.1 | PBR: True | Est. credits: 80 Task ID: a9d3e... [60s] Status: PROCESSING (28%) [180s] Status: PROCESSING (71%) [240s] Status: FINISHED (100%)

Done! pikachu.glb - 26.6 MB saved to workspace

Converted to STL and loaded in your slicer:

Pikachu 3D model loaded in Creality Print slicer, generated via OpenClaw

pikachu.stl - 25 MB ready for printing

Blender imported the GLB natively. Clean export, no issues. Ready to slice!

Type a message...

Object Generation

OpenClaw Chat

Direct gateway session

connected

Generate a medieval treasure chest with gold coins, a padlock, vines and moss. Use Hunyuan Pro with PBR.

Generating with Hunyuan Pro 3.1 + PBR...

Model: Hunyuan Pro | PBR: True | Est. credits: 80 Task ID: c4f8b... [90s] Status: PROCESSING (55%) [210s] Status: FINISHED (100%)

treasure_chest.glb - 14.2 MB saved to workspace

PBR model with metallic, roughness, and normal maps included. Ready for Unity, Unreal Engine, or Blender.

Now generate a perfect geometric vase

Done! Generated in ~2 minutes.

geometric_vase.glb - 8.6 MB geometric_vase.stl - 23.8 MB (print-ready)

Loaded in your slicer and ready to go:

Geometric vase STL loaded in Creality Print slicer, generated via OpenClaw

91.72 x 93.04 x 98.58 mm | 499,918 triangles

Clean hard-surface topology from Geometry mode. Ready to slice and print!

Type a message...

Image to 3D - upload a photo, get a model

Send an image to the chat and the agent converts it to a 3D model using TRELLIS.2 or Hunyuan.

Image to 3D with TRELLIS.2

OpenClaw Chat

Direct gateway session

connected

Convert this product photo to a 3D model with textures. Use TRELLIS.2 for speed.

Using TRELLIS.2 - fastest option for image-to-3D.

Model: TRELLIS.2 | Resolution: 1024 | Textures: True | Est. credits: 30 Encoding image: product_photo.png Task ID: e2b7a... [12s] Status: PROCESSING (40%) [24s] Status: PROCESSING (85%) [36s] Status: FINISHED (100%)

product_model.glb - 18.4 MB saved to workspace

Textured 3D model generated in 36 seconds. Includes base color, metallic, roughness, and normal texture maps. Ready for e-commerce, AR preview, or further editing in Blender.

Type a message...

Multiple ways to generate 3D models

The agent picks the best approach based on what you provide

Text to 3D

Describe any object in natural language. The agent uses Hunyuan Rapid or Pro to generate the full 3D model with textures from your description alone.

Image to 3D

Send a product photo, concept art, or reference image. The agent uses TRELLIS.2 or Hunyuan to reconstruct accurate 3D geometry matching your input.

Multi-View to 3D

Provide multiple angles of the same object for the most precise reconstruction. Hunyuan Pro supports front, side, back, and top views.

From chat message to 3D print

The complete pipeline: describe an object, generate it, convert to STL, and slice for printing. All from a chat window.

01

Describe

"Generate a geometric vase with clean topology for 3D printing"

02

Generate

AI creates GLB model with optimized geometry in ~2 minutes

03

Convert

Agent exports to STL (print-ready) and saves to workspace

04

Slice & Print

Load STL into Creality Print, Cura, or any slicer and print

STL conversion - the agent handles it

OpenClaw Chat

Direct gateway session

connected

Convert the pikachu GLB to STL for printing

pikachu.stl - 25 MB ready for printing!

Blender imported the GLB natively (it was a proper GLB, not a ZIP). Clean export, no issues. Ready to slice!

Type a message...

Three AI models, one skill

The 3D AI Studio skill gives the agent access to three different 3D generation models. It picks the best one for your request, or you can choose.

TRELLIS.2

Microsoft

Fastest & Cheapest

Image-to-3D only. Best for product shots, characters, and objects from photos. Generates in 25 seconds to 4 minutes.

Credits15-55 credits

Speed25s-4min

InputImage only

Hunyuan Rapid

Tencent

Balanced

Text or image input. Fast generation with good quality. Great for quick text-to-3D concepts and prototyping.

Credits35-55 credits

Speed1-2 min

InputText or image

Hunyuan Pro

Tencent

Highest Quality

Maximum detail with PBR textures, face count control, and generation modes (Normal, LowPoly, Geometry, Sketch).

Credits60-100 credits

Speed3-6 min

InputText, image, multi-view

Credit costs per generation

Pay-as-you-go with credits. No subscriptions. Failed generations are automatically refunded.

Use Case	Model	Credits	Speed
Quick concept from image	TRELLIS.2 (geo only)	15-25	25s-2min
Textured model from photo	TRELLIS.2 + textures	25-55	1-4min
Text to 3D (fast)	Hunyuan Rapid	35	1-2min
Text to 3D + PBR	Hunyuan Rapid + PBR	55	1-2min
Highest quality	Hunyuan Pro	60	3-6min
Game-ready PBR asset	Hunyuan Pro + PBR	80	3-6min
Max detail + multi-view	Hunyuan Pro (full)	100	3-6min

New users receive free credits to get started. Purchase credits →

Setup guide

Get from zero to generating 3D models in about 10 minutes

1

Install OpenClaw

terminal

npm install -g openclaw openclaw init

Requires Node.js 22 LTS or 24. The init wizard sets up your gateway and connects your AI provider.

2

Install the 3D AI Studio skill

terminal

clawhub install whale-professor/3daistudio-integration

Or browse the skill on ClawHub: whale-professor/3daistudio-integration

3

Set your API key

terminal

# Get your key from https://www.3daistudio.com/Platform/API export THREE_D_AI_STUDIO_API_KEY="your-api-key-here" # Windows (PowerShell) $env:THREE_D_AI_STUDIO_API_KEY = "your-api-key-here"

Get your API key here. New accounts receive free credits.

4

Start generating

terminal

# Start the gateway openclaw gateway start # Open http://127.0.0.1:18789/ or chat via Telegram/Discord # Then just say: "make me a 3D model of a medieval sword"

The agent handles everything: model selection, generation, polling, and downloading. Your GLB file appears in the workspace.

What you can build with OpenClaw + 3D AI Studio

Generate 3D assets for any workflow, directly from chat.

Game Development

Characters, props, and environments from text descriptions

3D Printing

Generate STL-ready models for FDM and resin printers

XR & Spatial

PBR assets for VR, AR, and spatial computing apps

E-Commerce

Product photos to interactive 3D models in seconds

About the 3D AI Studio skill

A community-built OpenClaw skill that wraps the full 3D AI Studio API

Skill Details

Name3daistudio-integration

Version1.0.3

RuntimePython 3.8+ (stdlib only)

LicenseMIT

DependenciesNone (uses urllib, argparse, json, base64)

AuthBearer token via THREE_D_AI_STUDIO_API_KEY

Available Commands

trellisImage to 3D with TRELLIS.2 (fastest, cheapest)

rapidText or image to 3D with Hunyuan Rapid

proText or image to 3D with Hunyuan Pro (highest quality)

balanceCheck your credit balance

statusCheck a generation task by ID

downloadDownload a finished model

Frequently asked questions

Start generating 3D models from chat

Install the skill, set your API key, and describe what you want. Free credits for new users.

Built by the community

The 3D AI Studio skill for OpenClaw was created by Whale Professor (@WhaleProfessor on Telegram). Thank you for building this integration and making 3D generation accessible from any chat interface.

Want maximum quality without OpenClaw?

Use Hunyuan3D 3.5 directly on 3D AI Studio with up to 8K PBR textures and 2M+ polygons. No installation needed.

Building a game?

Compare all AI 3D models for game development with engine compatibility and polygon budget guides.

Need the full API?

Access 3D generation, image generation, mesh tools, texturing, remeshing, and more via the REST API.