TL;DR: Meshy 6 is natively integrated into ComfyUI as an official Partner Node, so updating ComfyUI is all it takes, with no third-party plugin to install. Three official workflow templates ship in the Templates Library under the 3D category: text-to-3D, image-to-3D, and multi-view-to-3D. The Meshy node set covers the full asset pipeline inside ComfyUI, from generation through mesh refining, PBR texturing, auto-rigging, and animation to GLB export. Every workflow also runs on Comfy Cloud in the browser, so no local GPU is required.
Meshy 6 is now natively built into ComfyUI as an official Partner Node. If you use ComfyUI, you can generate production-ready 3D models from text, images, or multi-view references directly on your node canvas, then refine, texture, rig, animate, and export them without ever leaving your workflow. There is nothing to install. Update ComfyUI, open the Templates Library, and the Meshy 6 workflows are already there.
This guide covers what the integration includes, how it differs from third-party ComfyUI-Meshy plugins, and how to run your first 3D generation workflow, locally or on Comfy Cloud with zero GPU.
Quick Summary
| Integration type | Official Partner Node, natively built into ComfyUI core |
|---|---|
| Installation required | None. Just update ComfyUI to the latest version |
| Supported workflows | Text-to-3D, Image-to-3D, Multi-View-to-3D |
| Beyond generation | Refine, PBR texturing, auto-rigging, animation, 3D preview |
| Output format | Meshy nodes output GLB and FBX; templates can save GLB via SaveGLB / output directory |
| Local GPU needed | No. All Meshy 6 workflows run on Comfy Cloud |
| Official docs | ComfyUI Meshy 6 Partner Node documentation |
What Is the Meshy Partner Node in ComfyUI?
ComfyUI is one of the most widely used open-source, node-based interfaces for AI generation workflows, letting creators build repeatable pipelines for images, video, audio, and 3D. In January 2026, the ComfyUI team announced that Meshy 6 is natively available in ComfyUI as part of their Partner Node program.
Partner Nodes are a different category from community custom nodes. They are official API integrations maintained and distributed by the ComfyUI team itself, shipped inside ComfyUI core rather than installed through ComfyUI-Manager. For Meshy, that means the full Meshy 6 toolset arrives with a standard ComfyUI update: no plugin repository to clone, no dependencies to resolve, and no risk of the integration breaking when ComfyUI updates, because the same team ships both.
The integration is built on Meshy 6, the latest generation of our 3D foundation model, which brings refined geometry for characters and organic shapes, sharper hard-surface modeling for mechanical assets, a dedicated Low Poly Mode for real-time game assets, and improved structural accuracy across all model types.
How Is the Partner Node Different From Third-Party ComfyUI-Meshy Plugins?
If you have searched for "ComfyUI Meshy" before, you have probably found community-built extensions such as ComfyUI-Meshy on GitHub. Those plugins were a useful bridge, but the official Partner Node replaces them with something more complete and far lower-maintenance.
| Official Meshy Partner Node | Third-party ComfyUI-Meshy plugins | |
|---|---|---|
| Installation | None, built into ComfyUI core | Manual install via ComfyUI-Manager or git clone |
| Authentication | Log in with your ComfyUI account, no API key to manage | Paste your own Meshy API key into a node |
| Node coverage | Text-to-3D, Image-to-3D, Multi-View-to-3D, plus Refine, Texture, Rig, Animate | Typically Text-to-3D and Image-to-3D only |
| Maintenance | Updated by the ComfyUI team with each release | Depends on the individual maintainer |
| Comfy Cloud support | Yes, runs with no local GPU | Not available as cloud templates |
| Ready-made templates | Three official workflows in the Templates Library | Build your own graph from scratch |
The node coverage difference matters most in practice. The Partner Node is not just a generation endpoint. It exposes Meshy's full asset pipeline on the ComfyUI canvas:
- MeshyTextToModelNode / MeshyImageToModelNode / MeshyMultiImageToModelNode handle generation from each input type
- MeshyRefineNode upgrades a preview mesh with higher quality geometry and detail
- MeshyTextureNode applies PBR texturing, including metallic, roughness, and normal maps
- MeshyRigModelNode auto-rigs humanoid characters with clear body structure
- MeshyAnimateModelNode applies motion for quick previews and turntables
- Preview3D and SaveGLB let you inspect results in real time and export a GLB ready for any DCC tool or game engine
In other words, you can go from a text prompt to a textured, rigged, animated, exported asset inside a single ComfyUI graph.
What Can You Build With Meshy 6 in ComfyUI?
Three official workflow templates ship with the integration, one for each input mode. Each can be opened locally from the Templates Library or launched instantly on Comfy Cloud.
Text-to-3D: Generate Models From a Prompt
The Meshy 6 Text to Model workflow turns a written description into a 3D asset. Type a prompt, run the graph, and MeshyTextToModelNode generates the model. The template also bundles a Meshy Toolset group with Refine, Texture, Rig, and Animate nodes that you can enable to sharpen geometry and push the asset further. It is the fastest path for concept exploration, prop generation, and game-ready low-poly assets where you are starting from an idea rather than a reference. If you want to see what the underlying model can do before opening ComfyUI, the same engine powers Meshy's text-to-3D generator on the web.
Image-to-3D: Turn a Single Image Into a Mesh
The Meshy 6 Image to Model workflow converts a single 2D reference into a 3D mesh with UVs, with GLB and FBX outputs and optional texture and PBR settings. Load an image with LoadImage, connect it to MeshyImageToModelNode, then chain Refine and Texture nodes to iterate on quality. This is the natural fit for ComfyUI users in particular: generate a concept image with your favorite image model earlier in the same graph, then feed it straight into Meshy's image-to-3D engine to lift it into 3D, all in one pipeline.
Multi-View-to-3D: Reconstruct From Multiple Angles
The Meshy 6 Multi-Image to Model workflow accepts several views of the same subject and uses them together for more accurate proportions, better detail capture, and cleaner mesh structure. If you already have character sheets or front/side/back turnarounds, this mode gets you the closest match to your design.
How Do You Start Using Meshy in ComfyUI?
Getting your first Meshy 6 generation running locally takes four steps:
- Update ComfyUI to the latest version. Partner Nodes ship with core updates, so an outdated install is the most common reason the nodes would be missing. If you have updated and still do not see them, check whether you are on a Desktop or Cloud stable build, which can lag slightly behind the latest core release, or look for node import errors in the startup log.
- Open the Templates Library from the sidebar and browse to the 3D category.
- Select a Meshy 6 workflow: Text to Model, Image to Model, or Multi-Image to Model.
- Edit the prompt or swap in your image, then run the workflow. The finished model appears in the Preview3D node and is auto-saved to ComfyUI/output/.
For node-level parameters and requirements, the official ComfyUI documentation for the Meshy 6 Partner Node is the canonical reference and is kept current with every ComfyUI release.
Run Meshy 6 on Comfy Cloud With No Local GPU
Comfy Cloud removes the biggest barrier to node-based 3D: every Meshy 6 workflow runs in the browser with no local hardware. Because Meshy generation happens through the API rather than on your machine, you do not need a gaming GPU, a Python environment, or any local setup. Open the template link, edit the prompt, and run.
This makes the integration just as useful for newcomers who have never installed ComfyUI as it is for veterans with established local pipelines. If you have been curious about node-based 3D workflows but put off by the hardware requirements, Comfy Cloud plus Meshy 6 is the lowest-friction starting point available today.
Meshy 6 in ComfyUI brings together two things our community has been asking for: the control of node-based workflows and the quality of a production-grade 3D foundation model. Creators like kirinkarwai have already been pairing ComfyUI and Meshy in their daily pipelines; the Partner Node makes that combination native. Update ComfyUI or open a template on Comfy Cloud, and your first generated, textured, rigged asset is a few minutes away.
Want to push your models further? Bring any GLB from your ComfyUI output into Meshy's workspace to retexture, remesh, edit, and export to any format your pipeline needs.
Frequently Asked Questions
Do I need to install a plugin to use Meshy in ComfyUI?
No. Meshy 6 is an official Partner Node built directly into ComfyUI core. Updating ComfyUI to the latest version is all that is required. The workflow templates appear in the Templates Library under the 3D category.
Do I need a Meshy API key to use the Partner Node?
No Meshy API key is needed. The official Partner Node authenticates through your ComfyUI account login, and generations consume ComfyUI credits purchased in that account. Bringing your own Meshy API key is not supported at the moment.
How does pricing work for Meshy 6 in ComfyUI?
Partner Nodes run on ComfyUI's credit system: you buy credits in your ComfyUI account, and each Meshy 6 generation draws from that balance, so there is no separate Meshy subscription to manage. Credit costs are generally in line with what the same generation would cost through the Meshy API. If you want unlimited iteration inside the full Meshy editor, including retexturing and editing existing assets, you can also work directly in Meshy with a Meshy plan.
Can I use Meshy in ComfyUI without a GPU?
Yes. All three Meshy 6 workflows run on Comfy Cloud in the browser. Meshy generation itself is API/cloud-based, so no local GPU is needed for the Meshy nodes even when running ComfyUI locally.
What file formats does the Meshy node export?
Meshy nodes output GLB and FBX. In the official templates, SaveGLB nodes auto-save the results to ComfyUI/output/, and the Preview3D viewport includes an Export menu that can also convert the model to GLB, OBJ, or STL. GLB imports cleanly into Blender, Unity, Unreal Engine, Godot, and every major DCC tool. For other formats such as USDZ, you can convert the asset with Meshy's free online 3D file converter.
Can I rig and animate models inside ComfyUI?
Yes. The Partner Node set includes MeshyRigModelNode for automatic rigging and MeshyAnimateModelNode for applying motion, so eligible assets can come out of your graph already rigged and animated for previews or turntables. One caveat: auto-rigging currently works best on textured humanoid models with a clear limb and body structure. Untextured meshes, non-humanoid assets, and humanoids with ambiguous anatomy are not yet reliably supported.
What is the difference between the Meshy Partner Node and ComfyUI-Meshy on GitHub?
ComfyUI-Meshy on GitHub is a community-built extension that connects to Meshy through a user-supplied API key and covers text-to-3D and image-to-3D generation. The Partner Node is the official integration: it ships inside ComfyUI itself, adds multi-view generation plus the refine, texture, rig, and animate stages, includes ready-made templates, and runs on Comfy Cloud. For new projects, the Partner Node is the recommended path.
