Start with the subject and scene
Name the main subject, what it is doing, and the environment around it so the image has a clear visual premise.
Start with words only. Describe the subject, setting, composition, lighting, and visual direction you want, choose GPT Image 2.5 Flare or Sunburst, and generate a new image without uploading a reference.
Required Credits…
PROMPT-FIRST IMAGE CREATION
Name the main subject, what it is doing, and the environment around it so the image has a clear visual premise.
Add framing, camera angle, depth, lighting direction, time of day, or mood when those choices matter to the result.
Describe the material, rendering, photographic, illustrative, or design direction you want instead of stacking unrelated style words.
WRITE A BETTER IMAGE PROMPT
Lead with what must be in the image and what is happening. This establishes the core scene before decorative detail.
Use framing, viewpoint, focal emphasis, and lighting to control how the scene is presented rather than only what objects appear.
Finish with the visual treatment and any important constraints, such as clean negative space, realistic material detail, or a specific design mood.
HOW TEXT TO IMAGE WORKS
Describe the subject and scene first, then add composition, lighting, style, and only the constraints that materially change the result.
Start with Flare for faster iteration or choose Sunburst when precision matters more, then select the shape that fits the intended image.
Check the required Credit quote, generate once, and keep the finished image for download, reference-based editing, or Image to Video.
KEEP CREATING
Return to the broader Image workspace when you want the option to start from text or one reference image.
Open AI Image GeneratorCompare Flare and Sunburst and see the Image V1 settings currently available in Ikat.
Explore GPT Image 2.5Use an existing image as the reference when the job is transformation rather than creating from a blank starting point.
Edit an imageTake a finished still image into the motion workflow as the starting frame for a new video.
Animate an imageCreate images with GPT Image 2.5 and use the same Credit balance across Ikat's current video models and creation workflows.
See full pricingTEXT TO IMAGE FAQ
Text to Image creates a new image from a written prompt without requiring a source image. The prompt describes the visual you want the model to generate.
Start with the subject and setting. Add composition, camera or viewpoint, lighting, style, materials, and constraints only when they matter to the result.
Flare is the default for faster everyday iteration. Sunburst is available when tighter precision in detailed creative work matters more than speed.
Ikat Image V1 provides common square, landscape, and portrait aspect ratios while keeping the public output resolution fixed at 1K.
A successful GPT Image 2.5 Image V1 result costs 5 Ikat Credits. The quote is shown before Generate.
Yes. A completed result can be reused as the reference for another Image generation so you can describe the next change in plain language.
Yes. You can hand a completed image to Image to Video as the starting frame. The video request has its own settings and Credit quote.
Describe the visual, choose the GPT Image 2.5 path that fits the job, review the quote, and generate from text in Ikat.
Create from text