Kodo: chat-to-editable-design, layers included

What it is. Kodo is an AI design tool you talk to. You describe what you want — a poster, a slide deck, a menu, a social graphic — and it generates a layered design you can edit straight away. Text, colors, spacing, layout: all of it stays live and adjustable, the way a file from Figma or Illustrator would be, rather than a finished image dropped on your desk.

Why it matters. Most AI image tools give you a flat render. It looks good until you need to change the headline or swap a color, and then you are either back to prompting or opening the file in a separate editor and rebuilding things by hand. Kodo skips that detour by making the output editable from the start. For a designer who needs a quick first draft to show a client, or a social graphic that has to match a brand color, that difference is the whole point.

How to use it. Start with a prompt that includes the purpose, the tone, and anything specific about the content — “a menu for a Japanese ramen shop, minimal, dark background, three sections” works better than “a menu.” Kodo returns a structured design. From there you work directly in the interface, adjusting type, swapping colors, moving elements around. When it is close enough, you export. The realistic use case is getting from brief to a workable draft in a few minutes, then refining rather than building from scratch.

Limits. It is built for a specific set of formats — posters, slides, menus, social graphics — so if your work lives outside those, it may not fit. The designs it generates will reflect what the model knows about layout conventions, which means the output can feel generic until you push it with more specific prompts or manual edits. How well the export plays with other tools in your stack is worth testing before you rely on it for a real deadline.