Comfyg Queue

ComfygQueue is a custom node for ComfyUI designed to streamline batch image generation with multiple resolutions. Its boring to stay changing resolutions, to test which one is the best for each model, so I made this node for me too :)

Ignore the "ComfygQueueTrigger" node, I'm still working...

v1.2.0:

• New format output options: I'm running out of storage space, so I realized that the AVIF/Webp format is almost half the size of PNG, so I decided to make a new selection of options for the output file format: PNG, AVIF or WEBP. I'm still testing these formats and will think of more approaches to reduce size consumption.

---

Features

• Batch Resolution Support: Generate images at multiple resolutions in a single run.
• Seed Control:
  • Use a fixed seed or allow random generation.
  • Optionally increment +1 in seed for each resolution.
  • BUT! The node dont get the seed from the generated PNG, when you import, so pay attention.
• Latent Input Handling: If an input latent is provided, the node uses its resolution and skips multi-resolution loops.

---

Inputs

• Required:
  • res_presets: String, e.g., "1024x1024,1152x896"
  • save_prefix: Prefix for saved files
  • batch_seed: Integer, -1 for random, or a fixed seed
• Optional:
  • incremental_seed: Boolean, increment +1 per resolution
  • model, positive, negative, vae: Standard ComfyUI connections
  • latent: Optional latent image input
  • steps, cfg, sampler_name, scheduler, denoise: Sampling parameters
  • output_subdir: Subfolder for saved images
• Hidden:
  • prompt, extra_pnginfo

---

Outputs

• save_info: Summary of generation
• file_list: JSON array of saved file paths

---

Example Usage

1. Connect your model, prompt, and VAE as usual.
2. Set res_presets for desired resolutions.
3. Optionally provide a latent to preserve resolution.
4. Set batch_seed for reproducibility.
5. Run the node; files are saved in default output directory.

---

File Naming

<save_prefix>_<YYYYMMDD_HHMMSS>_<width>x<height>_<seed>.png

---

Contributing

Contributions and suggestions are welcome! If you encounter any issues or have ideas for improvements, please open an issue or submit a pull request.

License

This project is licensed under the MIT License.