AI image generator
Make and edit images by describing them
ϑbsidian writes art-directed prompts for GPT Image 2, so a simple request turns into a considered image. You watch it being painted in the chat, and it lands in your files, ready to download or to use in a website, deck or document it is building for you. Upload a photo to edit it: change the background, remove an object, restyle it, or combine several images into one.
How it works
- 1
Describe the image
A picture, a logo, a poster, a product shot, a comic. Say where it will be used and it picks the right shape.
- 2
Watch it appear
The image forms live in the chat, from a blurry preview to the final picture, then saves to your files.
- 3
Refine it
Ask for a targeted edit ("make the jacket red, keep everything else") instead of starting over.
Try asking
What you get
Strong prompts, written for you
Composition, style, lighting and exact text handled like an art director would.
Real photo editing
Edits that keep what matters (faces, framing, text) and change only what you ask.
Transparent backgrounds
Logos, stickers and product cut-outs as PNGs you can place anywhere.
Used in your projects
Generated images drop straight into the websites, decks and documents it builds.
Questions
Which image model does it use?
GPT Image 2 by OpenAI, which renders text in images well and preserves details when editing.
How many images can I make?
Each plan includes a monthly allowance of images, starting on the Free plan once your email is verified. The allowance resets on the 1st of each month.
Can it make vector logos or icons?
For crisp, scalable graphics it can also draw SVG by hand, which suits icons and simple logos better than a generated picture.