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 > FunUI > FunUI Checkpoint Loader

ComfyUI Node: FunUI Checkpoint Loader

Class Name

FunUICheckpointLoader

Category
FunUI/Loaders
Author
digital-garbage (Account age: 1276days)
Extension
FunUI
Latest Updated
2026-11-03
Github Stars
0.02K

How to Install FunUI

Install this extension via the ComfyUI Manager by searching for FunUI
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter FunUI 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

FunUI Checkpoint Loader Description

Loads diffusion checkpoints, detects model types, and warns of missing text encoders or VAEs.

FunUI Checkpoint Loader:

FunUICheckpointLoader loads a diffusion checkpoint so you can use its image-generation model, text encoder, and image-to-latent conversion components in a ComfyUI workflow. A checkpoint is a saved collection of trained model weights that helps determine the visual style and capabilities available for generation. The node asks ComfyUI to identify the checkpoint’s configuration automatically, so you do not need to select a separate configuration file. It also checks whether the checkpoint includes its text encoder and VAE, making incomplete checkpoints easier to recognize before they cause problems later in your workflow. Unlike a loader that simply loads the available components, it reports the detected model class and warns when important components are missing. It does not download missing components or repair incomplete checkpoints.

FunUI Checkpoint Loader Input Parameters:

ckpt_name

Required COMBO selection identifying the checkpoint to load. Its options come from ComfyUI’s registered checkpoint folders, normally including the models/checkpoints directory and any additional configured checkpoint locations. Selecting a different checkpoint changes the trained model and, when included in that file, its associated text encoder and VAE; this can change the available visual styles, generation capabilities, and compatibility requirements of your workflow. The selected filename is resolved to a file path before loading, and ComfyUI automatically attempts to identify its configuration. Options: the filenames returned by ComfyUI’s checkpoint registry, rather than a fixed list defined by this node. Default: no explicit default is specified in the node schema. Minimum and maximum: not applicable because this is a filename selection rather than a numeric setting.

FunUI Checkpoint Loader Output Parameters:

model

A MODEL object containing the loaded generation model, used to denoise latents during sampling. Latents are the compact working representation from which your final images are produced. Connect this output to a compatible sampler, or pass it through model-modification nodes before sampling. This output is not an image and does not perform generation by itself; it supplies the trained model that downstream nodes use. Its underlying model class is also reported in the status message, which can help you identify what ComfyUI recognized in the checkpoint.

clip

A CLIP object containing the checkpoint’s text encoder, when available. A text encoder translates your written prompts into information the generation model can use, so this output typically connects to text-encoding nodes for both positive and negative prompts. The output name and socket type are clip and CLIP, but the actual encoder configuration depends on the loaded checkpoint and ComfyUI’s support for it. If the checkpoint does not provide a text encoder, this output is None and the node reports clip as missing. Any downstream operation that requires a usable text encoder will need a separately loaded, compatible encoder instead.

vae

A VAE object used to convert between images and the compact latent representation used during generation. In a typical text-to-image workflow, it connects to a VAE decoding node to turn sampled latents into a visible image; in image-to-image workflows, it can also support encoding an existing image into latents. The VAE must be compatible with the selected model. If the checkpoint lacks a usable VAE, ComfyUI may return either None or an empty VAE placeholder that fails when used. This node checks for both cases and reports vae as missing, but it returns the original value rather than replacing the placeholder or loading a substitute.

status

A STRING summarizing the loaded filename and the detected underlying model class. Its normal format is filename loaded as ModelClass, where ModelClass is the class name identified from the loaded model, not necessarily a familiar checkpoint marketing name. If components are missing, the message adds (no clip), (no vae), or (no clip, vae). For example, the supplied test verifies unet_only.safetensors loaded as MiniMaxH3 (no clip, vae). This output helps you inspect what loaded and whether additional component loaders are needed; it does not guarantee that the rest of your workflow is compatible or that generation will succeed. If loading raises an exception, the node does not return a completed status message.

FunUI Checkpoint Loader Usage Tips:

  • Choose a checkpoint intended for your workflow’s model family. Automatic configuration detection removes the need to select a configuration file manually, but it does not make every checkpoint compatible with every sampler setup, text encoder, or latent-processing node.
  • For a straightforward workflow, use a checkpoint that includes the model, text encoder, and VAE. Connect model to your sampling path, clip to your positive and negative text-encoding nodes, and vae to your image-decoding node.
  • Check status after loading, especially when trying an unfamiliar file. A successful model load can still produce (no clip, vae), meaning the checkpoint is not a complete all-in-one package.
  • If a checkpoint intentionally contains only the generation model, supply a compatible text encoder and VAE through separate loaders. Connect those separately loaded components directly to the downstream nodes that need them rather than using the missing components from this loader.
  • Store checkpoint files in ComfyUI’s registered checkpoint locations. If a newly added file is not offered in ckpt_name, refresh the available model list or restart ComfyUI, and confirm that any custom model-folder configuration points to the correct directory.
  • There are no precision, device, or memory-management controls on this node. For memory-limited systems, choose a checkpoint appropriate for your available hardware and use ComfyUI’s broader memory settings rather than looking for additional loader parameters here.

FunUI Checkpoint Loader Common Errors and Solutions:

filename has no clip; those outputs are empty and whatever reads them will fail

  • Explanation: This is a warning produced when the checkpoint loads without a text encoder. The generation model may still be available, but prompt-encoding nodes cannot use the missing clip component.
  • Solution: Select a checkpoint that includes a text encoder, or load a compatible encoder separately and connect it to your prompt-encoding nodes. Check that the encoder matches the generation model’s requirements.

filename has no vae; those outputs are empty and whatever reads them will fail

  • Explanation: This warning means the checkpoint has no usable VAE. The returned value may be empty or may be a placeholder that only raises an error when an encoding or decoding node tries to use it.
  • Solution: Load a VAE compatible with your model and connect it directly to the relevant VAE encoding or decoding nodes, or choose a checkpoint that includes a usable VAE.

filename has no clip or vae; those outputs are empty and whatever reads them will fail

  • Explanation: The checkpoint lacks both a text encoder and a usable VAE. This can happen with a model-only file and does not necessarily mean that the model weights are damaged. The corresponding status suffix is (no clip, vae).
  • Solution: Provide both components through separate, compatible loaders, or switch to a complete checkpoint. Do not connect the missing components to downstream nodes and expect the loader to create replacements automatically.

Checkpoint file not found

  • Explanation: The selected ckpt_name cannot be resolved to an existing file in ComfyUI’s registered checkpoint locations. This can happen if the file was moved, renamed, deleted, or referenced by a workflow created on another machine. The exact exception wording comes from ComfyUI’s path-resolution code, not this node.
  • Solution: Confirm that the file exists in a registered checkpoint folder, refresh the available filenames, and select it again. If you use a custom checkpoint directory, check that ComfyUI is configured to search it.

Checkpoint loading or model-configuration detection fails

  • Explanation: ComfyUI’s underlying loader can fail if a file is damaged, incomplete, unsupported by your installed version, or not a suitable checkpoint. A similarly named file might instead contain a LoRA, a text encoder, or another component that requires a different loader. The supplied node code does not define a fixed message for these failures.
  • Solution: Read the accompanying ComfyUI error details, verify the file’s intended purpose and requirements, and use the appropriate loader. If the download is incomplete or corrupted, download it again from its trusted source. If the model requires newer support, check its documented ComfyUI version requirements.

Out of memory while loading

  • Explanation: Loading a large checkpoint can exceed available system memory or GPU memory. This is a possible underlying loading failure rather than a warning explicitly generated by FunUICheckpointLoader; the exact message depends on the runtime and hardware.
  • Solution: Close other memory-heavy applications, release models that are no longer needed, or choose a smaller compatible checkpoint. Review ComfyUI’s memory-management options if necessary; this node has no device or precision setting to adjust directly.

FunUI Checkpoint Loader Related Nodes

Go back to the extension to check out more related nodes.
FunUI
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.