LoadersPartial Capture

DualCLIPLoader

class_type: DualCLIPLoader

Source repo

Loads two CLIP text encoder models simultaneously, required for architectures like SDXL and Flux that use dual text encoders (CLIP-L and CLIP-G or T5) for richer prompt understanding.

What It Does

DualCLIPLoader addresses the dual-encoder architecture used by SDXL and Flux models. These architectures use two separate text encoders that process the prompt differently—typically CLIP-L (OpenAI) for general semantics and CLIP-G (OpenCLIP) or T5 for detailed understanding.

The node loads both encoder files and combines them into a single CLIP output that can be consumed by CLIPTextEncode. This is essential for workflows using modern model architectures that exceed the capabilities of a single CLIP encoder.

Inputs

clip_name1STRING

First CLIP model filename.

clip_name2STRING

Second CLIP model filename.

typeSTRING

Architecture type (sdxl, flux, sd3).

Outputs

CLIPCLIP

Combined dual CLIP encoder.

What Numonic Captures

  • Both CLIP model filenames
  • Architecture type

Known Gaps

  • CLIP model file hashes
  • Individual encoder versions

Related Nodes

CLIP Text Encode (Prompt)Conditioning
Load CheckpointLoaders
Back to ComfyUI Nodes

Capture ComfyUI metadata automatically

Numonic extracts workflow metadata from every ComfyUI generation — models, samplers, seeds, prompts, and custom nodes. Track provenance, maintain compliance, and never lose a workflow.

Return to the Guide