Describe it. Get the image.

Chat to create and refine avatars, posters, stickers, logos, illustrations, and photos. Your images stay in your browser.

Underlying AI models

Image generation currently uses GPT Image 2, Jimeng Image 4.0 (Seedream), and Grok Image 2. Users can view and choose the specific model in the selector at the top of Studio.

Conversational editing

After generating, describe changes in natural language. Studio refines the previous image instead of starting over.

Same capability as the API

Developers can call the same image capability through the OpenAI-compatible /v1/images/generations and /v1/images/edits endpoints.