Meshy AI 3D: Generate 3D Models

Lisa Ernst · 27.11.2025 · Technology · 6 min

In the last Gamejam, I used Meshy AI to generate stylized props for a scene within a few hours. This workflow, from text prompt to export, allowed for a complete set of small assets like market stalls and streetlights to be created without manually building each low-poly model.

Source: Own representation

Introduction to Meshy AI

Meshy AI is an AI-powered 3D model generator that transforms text and image prompts into complete 3D models with textures. The platform bundles modules such as Text-to-3D, Image-to-3D, AI Texturing for existing models, an animation module, and an API for integration into own applications.

Meshy generiert PBR materials with common maps such as Base Color, Roughness, Metallic, and Normal, which can be used directly in engines like Unity, Unreal, or DCC tools like Blender. The generated models can be exported in standard formats such as OBJ, FBX, GLB, USDZ, STL, or BLEND, facilitating further processing in typical 3D pipelines.

Meshy bietet einen free starter tier as well as several paid tiers which primarily differ in monthly credits and additional features. This makes the tool interesting for individuals, small studios, and larger teams, as entry is possible without high upfront investment.

For a quick overview of the interface and basic functions, a look at a compact video tutorial is recommended:

Workflow with Meshy AI

Before a prompt is entered into Meshy AI, it is helpful to define the model cleanly: intended use, level of detail, target engine, and style. For example, a low-poly prop for a mobile game requires less detail than a hero asset for a marketing visualization, even though both can be generated from text-to-3D.

For a clear start, a concise written description, supplemented by one or two reference images, is recommended. Meshy supports both pure text prompts and image-to-3D workflows, , where a photo or concept art serves as the basis.

A typical prompt for a game asset might look like this:

stylized, low-poly wooden market stall, canvas roof, a few crates with vegetables, game-ready, single 4K PBR material, clean topology, GLB export

Such prompts are based on the parameters that Meshy provides for style, topology, and material output, such as selecting the art style and the desired level of detail in the Text-to-3D module.

If, instead, you start from a product photo, you can upload the image in the Image-to-3D module and let Meshy automatically derive the basic shape, for example for a simple product visualization or a 3D print prototype.

Step 1: Text to 3D – Generating the Model in Meshy AI

In the dashboard, select the Text-to-3D feature and enter the prepared prompt into the input field. Then, choose a suitable model preset, depending on whether a character or a static asset is to be created.

Source: meshy.ai

Text-to-3D modeling in Meshy AI: Unicorn generation from a text prompt.

During generation, Meshy creates geometry and material data on the server side; the result appears in the 3D viewer of the web interface, where the model can be immediately inspected. The computation time depends on prompt complexity and output settings but is generally a matter of a few minutes, allowing for iterative improvements within a session.

Those who prefer to start with a guided video can follow a general Meshy AI tutorial to see how text and image prompts are converted into 3D characters, including initial material adjustments:

Step 2: AI Texturing and PBR Maps

Once the basic geometry is in place, texturing follows. Meshy offers its own AI Texturing module, which generates complete material sets from text prompts or reference images. The model is either passed directly from the Text-to-3D module to texturing, or an external mesh is uploaded.

In the texture interface, you describe material properties such as type of wood, degree of wear, or color palette, for example:

dark walnut wood, slightly worn edges, matte finish, subtle grain, for cozy fantasy tavern interior

Meshy erzeugt dabei PBR maps, suitable for real-time engines; in particular, Base Color, Roughness, Metallic, and Normal maps are highlighted to enable physically plausible light behavior. Third-party integrations like the Meshy-5-Retexture model on fal.ai also emphasize that high-quality PBR materials are generated.

Especially for assets placed close to the camera, a second or third iteration of the texture prompt is worthwhile. These iterations are economically justifiable due to the credit structure of the paid plans as long as the prompt remains focused.

Those who focus specifically on texturing can find practical examples on YouTube of how material prompts are formulated and applied to existing 3D meshes.

Step 3: Export to Blender and Unity

Once the model and material are consistent, export follows. Meshy allows downloads in several formats, e.g. OBJ, FBX, USDZ, GLB, STL, or BLEND, to transfer models to common DCC tools and engines.

For Blender, GLB is a good choice, as the format can bundle geometry, materials, and textures in one file, while FBX is often used in game pipelines. After importing into Blender, the model is checked in the Shader Editor; if necessary, nodes can be adjusted or textures reassigned, e.g., to change Normal Maps from 'Color' to 'Non-Color Data'.

Source: meshy.ai

Diverse Creations: Examples of 3D models generated with Meshy-2.

For Unity or Unreal, the exported formats are placed in the respective asset structure to build prefabs or blueprints upon them. Since Meshy specifically refers to real-time applications, the generated assets are geared towards typical game and real-time workflows.

Those who want to see the entire process from text prompt and image references to an animated asset can watch a comprehensive workflow video.

Practical Applications

In small indie teams, Meshy AI is primarily used to quickly generate props, simple characters, or blockout geometry that will be refined later in DCC tools. The Text-to-3D function is suitable for testing variants of a motif before selecting a final model and refining it in detail.

In agencies or product teams, Meshy can be used to create early 3D prototypes for presentations or internal reviews. The combination of Image-to-3D and AI Texturing allows for quick creation of simple 3D versions from existing product photos or sketches, which are sufficient for mockups and simple animations.

Source: user-added

3D models of an SUV, created with AI tools like Meshy AI, can be directly integrated into game engines like Unity.

For educational environments or university projects, it is relevant that Meshy, according to external reviews, explicitly addresses beginners and aims to facilitate entry into 3D modeling without requiring prior complex modeling skills. The official tutorials section offers a collection of step-by-step guides on Text-to-3D, Image-to-3D, Texturing, and Animation, which can be directly integrated into courses or internal training.

For all scenarios, the AI output does not serve as a final replacement for manual modeling, but rather as an accelerator in the early production stage. Especially in game productions, it remains advisable to optimize Meshy models after import, for example through manual retopology, LOD creation, or additional detail work, in order to achieve the performance goals and style specifications of the project.

Meshy AI shifts the balance between raw effort and fine-tuning in the daily work of 3D artists, game developers, and product teams. Where previously several hours were needed for a first rough model, today often a structured prompt plus a few iterations in the Text-to-3D- and Texturing module, is enough to get a solid starting point.

The key added value is not in automating complete productions, but in generating series of usable drafts in a short time, which are then refined with classic 3D expertise. With the export formats for Blender, Unity, and other tools, , the PBR materials, and the growing

tutorial base free plan can run through a first project and then decide based on credits, export needs, and team size whether an upgrade is worthwhile. Thus, Meshy AI becomes a tool that actually saves time in everyday work and complements 3D workflows with a fast, controllable AI component.