Modly logoModlyDownload Free

Modly 3D extensions

Every extension adds an AI model to Modly, so you can generate 3D from an image or a prompt. Model extensions handle reconstruction and generation; process extensions run post-workflow steps such as rigging, audio, or mesh cleanup. All of them install by pasting a GitHub repository URL in the app.

31

Modly 3D extensions published

2

Extension types: model & process

0

Cloud credits, everything local

Model extensions for Modly

Each one plugs a new generator into Modly, so you can drop in an image or a prompt and get a textured mesh.

SenseNova Vision 7B MoT

Model

Run SenseNova Vision 7B MoT locally for image understanding, segmentation, OCR, point clouds, and camera pose estimation.

@DrHepa · Sep 2026

View in the catalog →

PartCrafter

Model

Generate structured, separately editable 3D object parts from a single image, with optional background removal.

@DrHepa · Sep 2026

View in the catalog →

PESZzzz-Modly-hunyuan3d-2.1

Model

Hunyuan3D 2.1 patched for Windows with AMD GPUs and modest desktop PCs.

@PESZzzz · Aug 2026

View in the catalog →

TripoSG

Model

Flow-matching diffusion from a single image, with sharper geometry and finer details than TripoSR.

@lightningpixel · Aug 2026

View in the catalog →

Hunyuan3d mini Turbo

Model

Tencent's lightweight image-to-3D pipeline, tuned for fast turns on a small GPU.

@lightningpixel · Aug 2026

View in the catalog →

TRELLIS2 GGUF (with texture)

Model

Convert one image into a textured 3D mesh in a single pass using the Trellis.2 GGUF model.

@lightningpixel · Aug 2026

View in the catalog →

TripoSplat

Model

Single image to 3D with TripoSplat, wired into Modly as a standard image-to-mesh model.

@lightningpixel · Aug 2026

View in the catalog →

Hunyuan3d mini

Model

Hunyuan3D 2 Mini — the lightweight Tencent image-to-3D pipeline for everyday laptops.

@lightningpixel · Aug 2026

View in the catalog →

Hunyuan3D 2.1 Full - Low VRAM

Model

Native Hunyuan3D 2.1 Full image-to-mesh, optimized for low-VRAM NVIDIA GPUs with INT8, FP8 and FP16 inference.

AlefK1708 · Aug 2026

View in the catalog →

Pixal3D

Model

Turns a single input image into a textured GLB mesh with a compact local pipeline.

@DrHepa · Jul 2026

View in the catalog →

Trellis Text

Model

Generate textured 3D meshes straight from a text prompt using Microsoft's TRELLIS text models.

@DrHepa · Jul 2026

View in the catalog →

Shap-E

Model

OpenAI Shap-E text-to-3D generation with GLB mesh output and PLY sidecars.

@DrHepa · Jul 2026

View in the catalog →

Cube3d

Model

Research-only Cube3D INT4 text-to-mesh generation with UI-managed weights and GLB output.

@DrHepa · Jul 2026

View in the catalog →

DreamCube

Model

DreamCube RGB-D panorama and navigable scene generation with UI-managed weights.

@DrHepa · Jul 2026

View in the catalog →

Codex-Image (GPT image 2.0)

Model

AI image generation and image editing inside your Modly workflow.

@DrHepa · Jul 2026

View in the catalog →

Kimodo

Model

KIMODO motion generation, kinematic motion diffusion, and GLB rig retargeting workflows.

@DrHepa · Jul 2026

View in the catalog →

Pi3 and Pi3X

Model

Pi3 single-image and Pi3X multi-view 3D reconstruction with GLB/PLY point clouds and structured sidecar outputs.

@DrHepa · Jul 2026

View in the catalog →

Hunyuan3d- Part

Model

Hunyuan3D Part workflows for AI-powered 3D asset generation.

@DrHepa · Jul 2026

View in the catalog →

Wan22-TI2V-5B

Model

Wan2.2 TI2V 5B image-to-video extension for motion reference clips.

@DrHepa · Jul 2026

View in the catalog →

Process extensions for Modly

Post-generation steps — rigging, cleaning, audio, and external tooling — run as nodes inside a Modly workflow.

YUE2

Process

Local Python process extension for full-song generation, editable musical plans and ABC-based covers.

@DrHepa · Sep 2026

ACE-Step 1.5

Process

A highly efficient open-source music foundation model that brings commercial-grade music generation to consumer hardware.

@DrHepa · Sep 2026

AnyTop

Process

Prepare arbitrary BVH skeletons, generate or edit motion, and extract AnyTop DIFT correspondences inside Modly.

@DrHepa · Sep 2026

LATO.2

Process

Generate topology-aware meshes from input geometry with the full voxel, topology, VAE and T-Flow pipeline.

@DrHepa · Sep 2026

Qwen3-TTS CustomVoice

Process

Generate multilingual 24 kHz speech with the built-in CustomVoice speakers from Qwen3-TTS 1.7B.

@DrHepa · Sep 2026

Kokoro-ONNX

Process

CPU-only multilingual text-to-speech process extension using Kokoro ONNX.

@DrHepa · Aug 2026

modly-pymeshlab

Process

Mesh post-processing through PyMeshLab — clean, simplify, repair, and prepare meshes for the web, low-poly use, or 3D printing.

Lorchie · Jul 2026

modly-sprite-pipeline

Process

One extension, six nodes: turn a Modly 3D generation into game-ready pixel-art sprites.

Lorchie · Jul 2026

modly-comfyUI

Process

Run any ComfyUI workflow from inside a Modly workflow.

@Lorchie · Jul 2026

SkinTokens

Process

Mesh auto-rigging, skin weight generation, and rigged GLB workflow output.

@DrHepa · Jul 2026

UniRig

Process

Automatic rigging and character rig preparation for exported 3D models.

@DrHepa · Jul 2026

MOSS-SoundEffect

Process

MOSS-SoundEffect v2 text-to-audio sound effect generation with WAV workflow output.

@DrHepa · Jul 2026

Install a Modly extension in three steps

  1. 01 In the app, open Models and click Install from GitHub.
  2. 02 Paste the HTTPS URL of the extension repository, then confirm the installation.
  3. 03 Download the model or one of its variants, and use it from the Modly model list.

Ready to extend Modly?

Grab the free desktop app, install your first extension, and generate a 3D model offline — the whole loop stays on your machine.