Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > ComfyUI-SkinToken > SkinToken Rig

ComfyUI Node: SkinToken Rig

Class Name

SkinTokenRigTrimesh

Category
3D/SkinToken
Author
rizzlord (Account age: 2297days)
Extension
ComfyUI-SkinToken
Latest Updated
2026-07-16
Github Stars
0.05K

How to Install ComfyUI-SkinToken

Install this extension via the ComfyUI Manager by searching for ComfyUI-SkinToken
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-SkinToken in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

SkinToken Rig Description

Generates a skeleton and rig for a 3D mesh with SkinTokens, then exports it via Blender.

SkinToken Rig:

SkinTokenRigTrimesh uses a SkinTokens model to generate a skeleton and rigging for a 3D mesh, then exports the generated asset through Blender. It brings model-based rigging into a mesh workflow without requiring you to build the skeleton manually. You can keep the generated rig asset for later processing and use the exported file in a compatible 3D application. The node also supports an optional transfer-based export route and optional post-processing, although the supplied source does not define the visual effect of those operations.

SkinToken Rig Input Parameters:

trimesh

The 3D geometry to rig, supplied as a TRIMESH. The node copies a Trimesh mesh before processing; it can also combine mesh geometry from a Trimesh scene into one mesh. A scene without mesh geometry cannot be processed. There is no default value: you must connect a compatible mesh.

ckpt_name

The SkinTokens checkpoint to use for rig generation. Options come from available .ckpt files registered under skintoken; the node also includes its configured default TokenRig checkpoint in the choices. The default is DEFAULT_TOKENRIG_CKPT, whose filename is not shown in the supplied source. Checkpoint selection determines which model is loaded, so choose a checkpoint that is available or can be obtained by the node's model-asset setup.

device

Selects the requested device setting for model loading. Options are auto and cuda; the default is auto. Use auto unless you specifically want to request CUDA. The supplied source passes this choice to the model loader but does not specify exactly how automatic device selection works or what happens on a system without CUDA.

save_file

Controls where the export path is created. The default is True. When True, the node creates a numbered file path in ComfyUI's output directory; when False, it exports to a temporary directory. Both settings still produce an exported file, because the node reloads that file to create its mesh result. Choose True when you need a persistent export.

filename_prefix

The prefix used to construct the saved export path when save_file is True. The default is 3D/SkinToken_. It is passed to ComfyUI's output-path helper, which supplies the output folder, base filename, and counter. When save_file is False, the node instead uses a temporary filename, so this prefix does not determine that path.

file_format

The format of the exported rigged file. Options are glb and fbx; the default is glb. This choice sets the export file extension and is passed to the Blender export workflow through the output path. Pick the format your next 3D tool or workflow expects.

use_transfer

Selects the transfer-based export route instead of the standard asset export. The default is False. When enabled, export uses the generated asset's source mesh path as its transfer target; if that path is missing, export stops with an error. The supplied source does not establish that this route improves rigging quality, so use it when your workflow specifically calls for transfer-based export.

use_postprocess

Controls whether post-processing is applied to the generated asset after skeleton renaming and before export. The default is False. Enable it only when you want the implementation's post-processing step; its specific changes are not visible in the supplied source, so compare results before adopting it for a production workflow.

group_per_vertex

Sets the group_per_vertex value passed to the Blender export or transfer operation. The default is 4; the minimum is 1 and the maximum is 32. It is an export setting, not a model-generation setting. The supplied source does not show how the exporter uses the value, so keep the default unless you have a target workflow that requires a different setting.

bottom_center_origin

Sets the bottom_center_origin option passed to the Blender export or transfer operation. The default is False. Enable it when you want the exporter to use its bottom-center-origin behavior; it does not change the node's input mesh before rig generation. The exact coordinate adjustment is handled by export code not included here.

smooth_angle

Sets the smooth_angle value passed to the Blender export or transfer operation. The default is 55.0; the minimum is 0.0, the maximum is 180.0, and the interface step is 1.0. It is an export-side smoothing setting rather than a rig-generation control. Adjust it when evaluating the exported surface appearance, while checking the result in the application where you intend to use it.

skeleton_template

Selects a skeleton-label template applied to the generated asset before optional post-processing and export. The default is the label associated with SKELETON_TEMPLATE_KEEP. Choices come from SKELETON_TEMPLATE_LABEL_CHOICES, but their displayed names are not included in the supplied source. The setting affects skeleton naming; the source shown does not support a claim that it changes the generated joint layout.

top_k

A generation setting passed to the model that limits the number of candidate token choices considered at each sampling step. The default is 5; the minimum is 1 and the maximum is 200. Lower values constrain choices more tightly, while higher values allow more candidates. The source does not guarantee that a particular value will produce a better rig.

top_p

A generation sampling setting passed to the model. The default is 0.95; the minimum is 0.1, the maximum is 1.0, and the interface step is 0.01. It controls a probability-based pool of candidate token choices: lower values make that pool more selective. Change it gradually when comparing generated results.

temperature

A generation sampling setting that affects how strongly the model favors its more likely token choices. The default is 1.0; the minimum is 0.1, the maximum is 2.0, and the interface step is 0.1. Lower values generally make sampling more conservative; higher values allow more variation. The model is configured to sample, but the source does not promise a specific visual outcome from changing temperature.

repetition_penalty

A generation setting passed to the model to influence repeated token choices. The default is 2.0; the minimum is 0.5, the maximum is 3.0, and the interface step is 0.1. It affects the model's generation process rather than mesh geometry directly. Keep the default as a baseline and change it only when comparing generated rigs.

num_beams

The beam-count setting passed to the model's generation process. The default is 10; the minimum is 1 and the maximum is 20. Larger values can require more generation work, although the actual performance effect depends on the model implementation. The node also enables sampling, so do not assume that raising this value guarantees a more accurate rig.

SkinToken Rig Output Parameters:

trimesh

A TRIMESH loaded from the exported file after Blender completes the export. This lets you pass the resulting mesh geometry to another compatible node. It is not the original input mesh, and a plain mesh output should not be treated as proof that rig data survives in this socket; use the exported file or asset output when you need to inspect rigging.

rigged_path

A STRING containing the path to the exported glb or fbx file. Use this path to locate the file for inspection or downstream processing. Its location depends on save_file: a numbered output-directory path when enabled, or a temporary-directory path when disabled.

asset

The SKINTOKEN_ASSET produced by model generation, after skeleton renaming and any enabled post-processing. It is the generated asset passed into the export operation and is useful for downstream nodes that accept this asset type.

backend

A STRING identifying the Blender export route used: bpy when Blender's Python API is available as the selected backend, or the selected headless Blender backend otherwise. It describes how export ran, not the quality or format of the generated rig.

SkinToken Rig Usage Tips:

  • Start with the defaults and a valid mesh, then inspect the exported file in your target 3D application before changing generation settings. This gives you a baseline for judging whether a change helped.
  • Keep save_file enabled when you need to retain the rigged export. With it disabled, the file is written to a temporary path even though rigged_path is still returned.
  • Choose glb or fbx based on the next tool in your workflow. If rig information matters, inspect the exported file or use the asset output rather than relying on the trimesh output alone.
  • Leave use_transfer off for a standard export. Turn it on only when you intend to use the transfer route and can confirm the generated asset has a source mesh path.
  • When testing generation settings, change one of top_k, top_p, temperature, repetition_penalty, or num_beams at a time. Comparing exports is more informative than changing several settings together.
  • Adjust smooth_angle, bottom_center_origin, and group_per_vertex while evaluating export requirements; these values are passed to the exporter, not to the model-generation call.

SkinToken Rig Common Errors and Solutions:

Missing SkinToken model assets after automatic download attempt under ...

  • Explanation: At least one required asset is still missing after the node's automatic download attempt: the selected checkpoint, the default VAE checkpoint, or the Qwen directory. The message lists the missing paths.
  • Solution: Check the listed paths and make sure the required model assets can be downloaded or placed in the expected model locations. Then rerun the node.

Expected TRIMESH input, got ...

  • Explanation: The connected value is neither a Trimesh mesh nor a Trimesh scene that the node can convert.
  • Solution: Connect a node that produces a compatible TRIMESH value, and check that no other data type has been routed into this socket.

Scene does not contain any mesh geometry.

  • Explanation: The supplied scene has no mesh objects to combine into a single mesh.
  • Solution: Check the scene upstream and provide one containing actual mesh geometry.

Could not convert scene to a single trimesh.

  • Explanation: The scene conversion did not produce the single Trimesh mesh required for rig generation.
  • Solution: Prepare or export the source as a single compatible mesh before connecting it to this node.

SkinToken did not return a generated rig asset.

  • Explanation: Model execution returned a result, but that result did not contain the asset the node needs for export.
  • Solution: Confirm that the required model assets are present and that the input is a valid mesh. Retry with the default generation settings; if it continues, record the input mesh and full error log for diagnosis.

Transfer export requires the source mesh path on the generated asset.

  • Explanation: use_transfer is enabled, but the generated asset has no source mesh path for the transfer operation.
  • Solution: Turn off use_transfer to use the standard export route, or investigate why the source mesh path was not retained.

Blender export or exported-file loading failure

  • Explanation: The node must export a file through its selected Blender backend and then load that file as a mesh. Backend availability, export execution, file access, or file loading can therefore stop the node, but the supplied source does not establish one fixed error string for these failures.
  • Solution: Read the full error for the failing stage. Confirm that the Blender backend is available, the destination can be written, and the chosen glb or fbx export can be opened. Retry with the standard export settings to narrow down the cause.

SkinToken Rig Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-SkinToken
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

SkinToken Rig