Image generation
Found this helpful? Share it:
Found this helpful? Share it:
Describe an image and the assistant creates it for you. Every result is saved to your Drive automatically, so you can reuse it or share it with your team later 🎨
Image generation Free and upworks on every signed-in account, the free tier included.
Tell the assistant what you want to see. It works best when you are specific about style, composition, colors, and detail.
Instead of "a dog", try "a golden retriever sitting in a sunny meadow, watercolor style, soft warm tones". The more you describe, the closer the result matches what you had in mind. The image appears right in the conversation, and you can ask for tweaks or a completely different take.
idapt offers around 17 image models, and the assistant uses nano-banana by default. They range from very cheap, around $0.003 per image, up to roughly $0.167 for the most detailed models, so you can trade cost against quality. Name a model in your message to use a specific one.
Browse the full lineup on the Image models page to compare pricing and capabilities. Many models also do image-to-image: give one an existing picture plus a prompt and it transforms rather than starts from scratch.
Every generated image is saved to a Generated Images folder in your Drive, so you never lose one even after you close the chat. Find them in the Drive browser any time.
Want an image somewhere specific? Say so in your prompt, for example "Generate an app icon and save it to the assets folder". The assistant creates the folder if it does not exist yet.
Point the assistant at an existing image for style or inspiration.
Reference a file with @in your message to attach it,
then ask for something in the same style. See Using @mentions and references.
Better prompts lead to better images. A few strategies:
Be specific about style: "oil painting", "pixel art", "photorealistic", "minimalist line drawing".
Describe the composition: "close-up", "aerial view", "centered", "rule of thirds".
Name colors and mood: "warm sunset tones", "dark and moody", "bright and cheerful".
Add detail cues: "high detail", "simple", "with a blurred background".
A vague prompt like "sunset picture" works, but "oil painting of a sunset over mountains, warm colors, high detail, dramatic clouds" produces a much better result.
Generate a few variations of the same idea, then refine the one you like best with a follow-up prompt.
Related articles
Was this helpful?