ComfyUI-Automation

ComfyUI-Automation

A collection of custom nodes for ComfyUI designed to facilitate complex looping and iteration workflows, specialized in media batch processing (Image + Audio + Video).

Overview

The core feature of this suite is the Stateful Iterator System, which allows you to iterate through a list of items across multiple generations without causing circular dependency errors in ComfyUI.

Nodes

1. Iterator Item

The basic building block. It gathers media components into a single bundled object.

• Inputs:
  • image: The source image for this item.
  • audio_text: Multiline string for audio generation (e.g., TTS).
  • video_prompt: Multiline string for CLIP text encoding or video generation.
• Outputs:
  • IMAGE, AUDIO_TEXT, VIDEO_PROMPT: Individual passthrough components.
  • ITERATOR_ITEM: The bundled object used by the Iterator List.

2. Iterator List

The controller that manages which item is currently active.

• Inputs:
  • iterator_id: A unique string identifier to sync state with the Iterator Signal.
  • reset: When enabled, forces the index back to 0 (the first item).
  • item1 to item6: Slots for your ITERATOR_ITEM bundles.
  • trigger: (Optional) A dummy input (accepts any type) used to force this node to execute after some other node in ComfyUI (useful for manual execution ordering).
• Outputs:
  • IMAGE, AUDIO_TEXT, VIDEO_PROMPT: Standard types from the currently active item.
  • IS_FINISHED: Boolean that turns True when the last item in the list is reached.

3. Iterator Signal

The "trigger" that tells the system to advance.

• Inputs:
  • image: The final output image of your processing chain.
  • iterator_id: Important: Must match the iterator_id in the corresponding Iterator List.
  • active: Toggle to enable/disable the advancement logic.
• Outputs:
  • SIGNAL: Boolean trigger status.
  • IMAGE: The input image (passthrough). This allows you to chain a Save Image or Preview node after the signal to ensure the loop captures the result.
• Logic: When this node receives a valid image, it increments the index for the specified iterator_id. If IS_FINISHED is False, it automatically queues the next iteration.

4. Iterator Counter

A simple numeric counter that increments with each iteration step.

• Inputs:
  • iterator_id: A unique string identifier to sync state with the Iterator Signal.
  • start: The starting integer value.
  • max_iterations: The total number of iterations to run.
  • step: The amount to increment by each step.
  • reset: When enabled, forces the counter back to the start value.
• Outputs:
  • current_count: The current integer value (start + index * step).
  • is_finished: Boolean that turns True when the max iterations are reached.

5. Video Concatenation

A specialized node that joins multiple video files into a single sequence, with optional transitions.

• Inputs:
  • transition_type: The type of transition to use between clips (e.g., fade, slideleft). Default is none.
  • transition_time: Duration of the transition in seconds.
  • output_format: The format of the concatenated video (e.g., mp4, mkv, gif).
  • video1 to video5: Input video objects (supports paths, lists, or ComfyUI video objects).
• Logic:
  • Concatenates videos sequentially.
  • Applies FFmpeg xfade transitions if selected.
  • Saves the result to the ComfyUI temp directory.
• Outputs:
  • video: The concatenated video object (compatible with ComfyUI video nodes).

How to setup a Video Batch Loop

1. Create several Iterator Item nodes with your source images and prompts.
2. Connect them into Iterator List. Give it a iterator_id like my_sequence.
3. Build your workflow (e.g., Upscale -> KSampler -> Video Combine).
4. At the very end, connect your generated image to a Iterator Signal. Give it the same iterator_id (my_sequence).
5. Use Auto Queue in ComfyUI.
6. Each time a generation is completed, the system advances to the next item.
7. Once your batch is done, connect the output paths of your generated videos to the Video Concatenation node to create the final unified video.

Examples

I have included example workflows in the examples/ directory:

• basic_image_loop.json: A simple demonstration of iterating through two images.
• video_batch_loop.json: A more advanced workflow showing video generation and automatic merging.

Installation

1. Navigate to ComfyUI/custom_nodes.

2. Clone this repository:

   git clone https://github.com/your-repo/ComfyUI-Automation.git
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Restart ComfyUI.