Hi OpenAI team and community,

I would like to request an official OpenAI-supported ChatGPT and Codex integration for ComfyUI.

ComfyUI is widely used for complex image and video generation workflows, but workflow building can become difficult when dealing with custom nodes, model loaders, CLIP/VAE issues, batch generation, prompt templates, workflow JSON, and automation scripts.

A built-in or officially supported OpenAI integration would be very useful for creators who need reliable help inside ComfyUI.

Suggested ChatGPT features inside ComfyUI:

  • Explain nodes and workflow errors
  • Review workflow structure
  • Help troubleshoot missing models, loaders, CLIP, VAE, and custom node issues
  • Generate prompt templates
  • Help design image/video workflows
  • Help with CSV or data-driven prompt batching

Suggested Codex features inside ComfyUI:

  • Generate helper scripts for batch workflows
  • Create workflow JSON automation tools
  • Build CSV-to-prompt automation
  • Help save outputs with structured filenames
  • Review workflow-related Python scripts
  • Support safe local file handling without unexpectedly modifying the ComfyUI install

Example use case:

I want to generate one image per player for a full sports roster. Each image should use different player data, keep a consistent visual style, and save with a useful filename. ChatGPT could help design the workflow, while Codex could create the helper scripts.

This would be especially helpful because third-party ComfyUI assistant tools can depend on hosted MCP/cloud servers that may go offline or return service errors. An official OpenAI-supported option would be more reliable for serious creators.

I understand that Codex is currently available through Codex web, app, CLI, and IDE extension, and that a community extension using the OpenAI API may be possible today. Still, I would like to request an official supported ComfyUI integration as a long-term product feature.

I would also be interested in hearing whether others are working on safe ComfyUI extensions using the OpenAI API, especially ones that avoid exposing API keys in frontend/client code.

Thank you for considering this.