OpenAI launched the heavyweight GPT-Image-1 API (the same image model as ChatGPT 4o) — now call it directly through ComfyUI's native API node (beta)! No complex configuration, no ChatGPT account needed; supports WeChat/Alipay top-up — easily use OpenAI GPT-Image-1, DALL·E 3, and DALL·E 2, the three top image-generation models.
\The nanny-level tutorial: ComfyUI workflow batch image generation
\1. Environment prep
– Make sure you've installed the latest ComfyUI (supporting API nodes)
– Register an OpenAI API account (only email needed, no ChatGPT Plus)
– Top up API credit via WeChat/Alipay (supports small-amount payments, flexible use)
2. Model-call configuration
1. In the ComfyUI workflow, add the OpenAI Image Generation node
2. Choose a model (GPT-Image-1 / DALL·E 3 / DALL·E 2)
3. Enter a prompt (supports Chinese descriptions, auto-optimizes generation results)
4. Set parameters (resolution, generation count, style modifiers, etc.)
3. Batch image-generation techniques
– Workflow automation: combine with the Batch Processing node to generate multiple image groups at once
– Style control: specify art style via the "in style of..." keyword (e.g. anime, realism, oil painting, etc.)
– Optimize prompts: use ChatGPT to assist prompt optimization, making image generation more precise
4. Payment and cost control
– Pay-per-use: DALL·E 3 is about $0.04 per image, suited to small-scale testing
– Top up directly via WeChat/Alipay, no overseas payment method needed
– Supports API-usage monitoring, avoiding over-spending
Summary
Now you can call OpenAI image models directly through ComfyUI — more flexible than Midjourney and more worry-free than Stable Diffusion! Whether for e-commerce assets, concept design, or social-media content, you can quickly generate high-quality images. Hurry and try it! 🚀
(Attached: the complete workflow JSON config file and example prompt templates are available via DM.)