gpt-image-2

You generate brand-safe images with embedded text and multilingual typography by calling the RunComfy endpoint for text-to-image with a prompt and a supported fixed size. This ensures accurate signage and brand marks for production workflows.

Yes, you can perform image edits using up to 10 reference image URLs for guidance and size control. The process invokes the RunComfy edit endpoint with your reference images, natural-language edit instructions, and size constraints over HTTPS.

The best way to edit product packaging while preserving details is using the RunComfy image editing endpoint with natural-language instructions. It follows your directives to keep specified elements intact while applying the requested modifications.

GPT Image 2 supports only specific fixed sizes for text-to-image generation to keep outputs consistent for production workflows. You must select from the supported size constraints when invoking the RunComfy endpoint.

Choose GPT Image 2 when directive precision for e-commerce, ad UI mockups, packaging, and posters is critical, especially when accurate text and logos matter. Route to sibling model skills if you explicitly need Flux 2, Nano Banana Pro, or Seedream.

Text-to-image generation requires specific size constraints to maintain output consistency for production workflows. The RunComfy API enforces these fixed dimensions to ensure reliable layout details and brand-safe visual results.