Use Agentic AI
Describe what you want, and Agentic AI picks the model, prompt, and settings for you.
Agentic AI is Kittl's new way to generate with AI. Instead of picking a model, writing a detailed prompt, and configuring settings yourself, you just describe what you want, and Agentic AI figures out the rest, including the model, prompt, style, and format. Use it any time you want a strong first result without needing to know how the AI stack works.
Generate with Agentic AI
Open a project or start a new one.
In the AI input at the bottom of the canvas, click the mode selector and choose Agentic AI.
Choose a starting point: a blank canvas, an image, a product photo, or a logo.
Describe what you want to create in plain, simple words. There's no need to write an overly detailed prompt.
Click Generate.
Agentic AI can create up to four variations of your idea at once, exploring different layouts, styles, and compositions. Compare the results and keep the one that works best.
How tokens work
Agentic AI uses tokens from your plan to generate images and videos. Since it picks the best model for your request automatically, the exact token cost depends on your output. As a general guide:
Images: approximately 20 tokens per image
Videos: approximately 35 tokens per second of video
Example prompts to try
Try describing your idea like this:
Designs: "Create three retro-style t-shirt designs about coffee."
Product imagery: "Place my product on a marble kitchen counter with soft morning light."
Ads: "Create a square Instagram ad for my skincare product, aimed at young professionals."
Social media videos: "Turn my product photo into a 10-second Reel with subtle movement."
Troubleshooting
I don't see Agentic AI in my editor.
Check the mode selector inside the AI input box. If it shows Generate Image or Generate Video instead of Agentic AI, click it and select Agentic AI from the list.
I got several images instead of one.
That's expected. If your prompt implies more than one result, like "four dog stickers" or "different layout options," Agentic AI generates separate outputs for each rather than combining them into a single image.
My aspect ratio isn't what I expected.
Agentic AI picks the aspect ratio based on your prompt. Mention the platform or format you're designing for, like "square Instagram post" or "vertical Reel," so it can match it more closely.
Common questions
Can I still write a detailed prompt myself?
Yes. Agentic AI only rewrites prompts that need more detail. If your prompt is already specific, it uses your wording as is.
Does Agentic AI work for video as well as images?
Yes. It picks the right model for the task, whether that's an image or a video, including turning a static design into a video with movement.
Can I run a multi-step or sequenced generation with Agentic AI?
Not yet. Sequenced generations aren't supported in this version.
Did you remove the regular image and video generator?
No, they're still there. Use the mode selector inside the AI input box to switch to Generate Image or Generate Video whenever you want that instead.