Introduction
Not every 3D asset begins as a creation of a 3D artist. Some begin with sketches, product photos, verbal descriptions, or ideas. The problem has been to take that starting content and convert it into editable geometry that can be controlled and manipulated in various ways to become a final asset. AI 3D generation has altered this process by pushing some early modeling steps out of software to generative systems.
One of the early AI 3D model generation platforms is Meshy, which uses AI for 3D model generation. While AI will not replace 3D software, it can generate a starting model that can be reviewed for discrepancies and edited to a larger degree prior to the intended use for the asset.
Meshy is one of the pioneer AI 3D model generation tools.
What Is Meshy?
Meshy is a web platform for generating and processing 3D content using AI. It has Text to 3D and Meshy image-to-3D functionalities that give users the flexibility to use modeling in a written or visual way, respectively.
That distinction is important because the best input differs by project. A game developer may know precisely how to describe an object, while a designer may have a concept image. A 3D printing enthusiast may have a picture of an object and wants to study a digital approximation rather than build a model.
The platform uses AI to assist with texturing, rigging, and image generation, among other things. Builders of custom workflows can employ Meshy, which exposes its 3D generation capabilities via an API and thus can be integrated into applications of automated production.
How Does AI 3D Generation Work?
AI 3D generation uses a simple approach. Designers provide a system with a description of a target object. Then, trained models suggest a 3D structure of the target object.
The approach is more interesting when the input is an image. A single image captures only a single view of an object. So, a model has to fill in the gaps of what the unseen portions looks like. Multiple images can help.
In such cases, Meshy’s multi-image input can be useful. Rather than purely relying on a single view of an object, users can supply multiple images of an object to provide more shape and appearance data.
Text to 3D
Text to 3D starts with a written description.
One prompt could be a stylized medieval shield or sci-fi case. A differnt example could be a cartoon styled vehicle. In all these cases, the generated images or modeling can be put directly into Digital Asset Management (DAM) systems for use. The reasoning behind this is artists aren’t expected to create each modeling polygon.
Text generation is perfect for the ideation phase where creators need to test several combinations.
Image To 3D
This approach is more like the opposite of text to image. To keep this simple for creators, the walkthrough simplifies the process, but in this case, the creator has to supply the system with a visual input.
Meshy offers a service where creators can develop a work-flow with use of 2D images and start with a 3D model.
The more angles used to describe and define a 3D model, the more effective meshy is.
It is especially helpful for detail that will skew the form of your requested model.
A Practical AI 3D Workflow
Stage 1. Creating A Starting Point
Consider if the object is better described with text or images.
If you don’t have a list of photographs and looking to describe the object you need, text is the way to go. If you can, describe the object with multiple photographs. This will help avoid a situation where the AI has to guess.
Stage 2. Generating a Rough Model
Send the description or the images you made and wait for the system to construct a model.
The first result should usually be treated as a starting point. AI-generated geometry has flexibility in topology, proportions, detail, and other characteristics, so inspection is important.
3. Check the geometry
Check the model from different angles.
Be sure that you have checked everything, including the bottom, back and side, thin and connected elements of the model, and everything that may have been buried in other model elements.
In both game design and 3D printing, additional consideration needs to be given to polygon density and topology, as well as watertightness and dimensions respectively.
4. Enhance aesthetics
Meshy’s AI texturing will likely help transform basic geometry into an asset with a more developed visual appearance.
This can help quickly visualize basic geometry for prototyping and concept work, but heavy adjustments to materials will still be required for production work to meet a given art direction.
5. Export the asset
Meshy makes it easy to move the assets generated in their system to other tools.
They allow common 3D formats (STL, OBJ, GLB, FBX), which increases their usability and functionality.
Where Can AI-Generated 3D Models Be Used?
AI-generated models will fit into many different processes and workflows, but the amount of work required depends on how final the models need to be.
Game Development
Game teams can get a lot of use out of AI-generated models with prototype work, and can use them during the conceptual and experimental stages of design.
Take the developer working on an early game level – they will generate several different environmental props and see which direction generates the most interest.
Something important to note is that, while an AI might have the ability to generate a three dimensional model, it is not ready to be imported into the engine or production pipeline in its current state. An artist would still need to optimize the geometry, generate UV’s, apply materials, and ensure the asset meets technical requirements for the game/project.
3D Printing
Ai model generation tools have a clear utility for 3D printing, given that most tools can export STL files.
A user is able to generate an idea and produce a model that they then evaluate the quality of and prepare for digital slicing prior to the physical print. 3D printing introduces requirements not found in traditional digital modeling, so it is important to verify printability.
Product and Industrial Design
Designers can use generative 3D tools at the early stages of a project to explore ideas at a higher level prior to investing the time to detail complex models.
It is useful especially at the idea generation stage of a project to understand what is possible and not to simply develop the final model.
Creative and Visual Design
AI tools for geometry generation can also be used by artists and hobbyists to generate ideas quickly.
Instead of building every single idea, users are able to generate several possibilities from which they can iterate.
Meshy’s Browser Based Tool Kit
Not all 3D processes require something as robust as a full blown modeling tool. Meshy includes a free browser based 3D Tool Kit that includes converters, viewers, STL repairs, and more.
This is valuable because bringing together common file-handling tasks and AI file generation within the same workspace simplifies file management.
Consider a user working with an STL file. They might have to check on the file, or even fix it, before they can continue to the next process in the workflow. utils that run inside a browser are far easier to use than installing a separate program to handle such a simple task.
Using a fully browser-based utility also lowers the initial barrier to the people comfortably using desktop-based 3D software.
Is Meshy Free to Use?
Meshy has a free plan for users looking to toy around with its AI for generating 3D structures.
Free tiers become especially important when dealing with AI-driven tools. Each tool’s offerings can be unpredictable at times, and may require adaptation within a workflow to fit your needs. It can be worth your while to test the capabilities of the tool before integrating it into your full workflow.
Meshy’s current plans and usage limits do fluctuate, so a potential user should know those to avoid service disruptions.
Meshy for Developers
Meshy isn’t the only service out there establishing a workflow for developers focused on AI-driven 3D generation.
There are now frameworks and APIs that developers can integrate to generate 3D assets without having to jump in and out of a web interface.
Automated pipelines can be established for creative tools, 3D asset generation, and generative interfaces.
The API approach is recommended when 3D generation is only one part of a multistep bigger process.
What Should You Check Before Using an AI-Generated Model?
AI will speed up a given step of the process involved in 3D model creation. However, the generated models should not be viewed as completed production models.
A model should be checked for the following:
Geometry: Distorted or disconnected components should be checked and removed as they should not be present. Holes should also be checked.
Topology: The mesh should be checked to ensure it fits the given purpose.
Scale: 3D printed models should be checked for dimension to ensure feasibility.
Textures: Check resolution and ensure seamlessness, material and consistency.
Optimization: Geometry should be reduced to improve performance if the given model needs to perform real-time tasks.
Rigging: Verify generated or applied rigs for assets that will be animated.
Licensing and project requirements: Ensure the model falls within the legal bounds of the project requirements.
Frequently Asked Questions
Is Meshy an AI 3D model generator?
Meshy is a platform that is based on AI for generation of 3D content. It supports both text-to-3D and image-to-3D workflows as well as other AI assisted features.
Can Meshy generate 3D models from multiple images?
Multiple images can be inputted to Meshy, allowing for the system to have more views of the same object and therefore having more information for the 3D model reconstruction.
What formats can Meshy export?
Meshy is able to export to STL, OBJ, GLB, and FBX formats, depending on the required optimization for printing, games, design, animation, etc.
Does Meshy have a free version?
Meshy has a free tier which lets users try its platform and AI 3D workflows.
Can developers integrate Meshy with their applications?
Meshy has an API for developers which means developers can integrate Meshy with their software and automate workflows.
Is Meshy fully browser based?
Meshy is a browser-based platform; users do not need to have traditional 3D software installed to use Meshy’s generation tools and related services.
Final Thoughts
The most important thing about 3D AI generation is lessening the gap between a thought and a working digital artifact. Meshy has a few different options for making this gap smaller, including text prompts, image prompts, and multi-image prompts. In addition to this, Meshy contains AI texturing, auto rigging, image generation, toolkit functionalities, developer APIs, and exports.
Meshy also has a unique user experience. In addition to AI generation, Meshy offers a free Tool Kit which contains conversion, online viewing, and STL repair tools. This looks to extend the platform’s usage beyond those creating new 3D models to those that use 3D models.
The goal is to make AI generation a small steps task in a big workflow. A generated model will save time during concepting or creating an initial asset, however inspection, editing, and preparation of the model for the overall workflow will still be required. For creators that are willing to integrate AI generation with common 3D generation practices, Meshy is a great option to try making previously time-consuming parts of digital creation less time-consuming.
