About this article: it is an overview based on the tools' public product pages and documentation, not a hands-on benchmark. We did not score or rank the tools. Features and prices change often, so check each vendor's site before deciding. 3DView.ai is our own product.
What a text-to-3D generator does
A text-to-3D generator turns a written description ("a wooden treasure chest with iron bands") into a textured 3D mesh. Most tools also accept a photo or a drawing as input (image-to-3D), which usually gives more control over the shape. The results work well for props, product mock-ups and concept work; they are rarely ready for close-up hero assets, animation or manufacturing without cleanup.
Questions to ask before choosing
- Inputs: text only, a single image, or several photos of the same object?
- Outputs: which file formats can you download (GLB, FBX, OBJ, USDZ, STL)?
- Integration: is there an API, or plugins for Blender, Unity or Unreal?
- Pricing model: subscription, pay-as-you-go credits, or open model weights you run yourself?
- License: what may you do with the models commercially? Read the terms.
The main tools
Meshy
A web app for text-to-3D and image-to-3D, plus AI texturing of existing meshes. It offers downloads in common formats (such as GLB, FBX, OBJ and USDZ), an API and plugins for popular 3D tools. Pricing is subscription-based, with a limited free tier.
Tripo
Text-to-3D, image-to-3D and multi-view-to-3D generation, with an API for developers and extra tools such as automatic rigging. Pricing is based on credits and subscriptions.
Hunyuan3D (Tencent)
An open-source family of image-to-3D models. The code and weights are published, so technical teams can run it on their own GPUs; a hosted demo is also available. Check the license terms for your region and use case.
Stability AI 3D models
Stability AI has released image-to-3D models such as Stable Fast 3D and SPAR3D, which build a textured mesh from a single image in seconds on suitable hardware. They are available as model weights under Stability's license and through its API.
Luma AI
Luma became known for capturing real objects and places from phone video (NeRF and Gaussian splatting) and released a text-to-3D tool called Genie. Its recent focus has been AI video, so check what it currently offers for 3D.
3DView.ai
3DView.ai does not train its own 3D model: generations run on third-party generation models, and 3DView.ai provides the workflow around them. What it offers today:
- Text-to-3D and image-to-3D from one photo or several photos (10 credits per model).
- Download as GLB, and an in-browser viewer with AR on supported phones.
- A shop of ready-made models from Sketchfab (10 credits, hosting included).
- Hosting with an AR link, a QR code and an embeddable viewer, included with no monthly fee.
- A REST API, available after your first credit pack purchase.
- One-time credit packs, no subscription: Test (80 credits, $7), Standard (250 credits, $19), Pro (750 credits, $49). Credits never expire. New verified accounts get one free AI model from text.
It does not offer mesh editing, rigging or FBX export; for those, open the GLB in Blender.
How to pick
- You need many assets through an API, or a plugin inside your 3D software: look at the tools that offer those integrations.
- You have your own GPUs and technical staff: open models such as Hunyuan3D or Stable Fast 3D avoid per-model fees.
- You want a model plus a shareable AR link or an embed for a website, paid per use: that is what 3DView.ai is built for.
Whatever you choose, test with your own objects: quality depends a lot on the subject and on the input image.