CookooChat
One app for every kind of generation.
Overview
Most people now have four AI apps on their phone: one for chat, one for images, one for voice, one for video. CookooChat's pitch is that they shouldn't have to. It's a single assistant where a conversation can produce a paragraph, a picture, a soundtrack or a clip without ever leaving the thread.
Text, image, audio and video in one assistant
Swap chat models without leaving the conversation
Credit model with a visible starter grant
The challenge
Bundling four generation modes into one product usually produces a mess — four disconnected tools wearing the same skin. The interface had to make switching between text, image, audio and video feel like one continuous surface rather than four apps stapled together.
There was also a business problem to design around. Generation costs real money per call, so the app needed a credit system that people could understand at a glance and that never left them guessing why a balance dropped.
Inside the app
A few screens from CookooChat. Tap any of them to take a closer look.
How we approached it
We built the chat as the home base and hung the other modes off it as intent shortcuts — Create Image, Generate Audio, Generate Video — sitting right above the input. You start typing; if what you want is a picture, one tap reframes the same conversation into image mode. Nothing about your place in the thread is lost.
Model choice is exposed but never in the way: a quiet selector at the top of the chat lets power users swap models mid-conversation while everyone else ignores it. Multimodal input — dropping in an image and asking about it — falls out of the same box people already type into.
The credit system got a deliberately friendly first run: a visible free grant on day one, a coin counter that's always on screen, and generation states that show exactly when a credit is being spent. Money in an AI app is anxiety, and the design's job was to lower it.
What we built
Four modes, one thread
Text, image, audio and video generation live inside a single conversation.
Model switching
Change the underlying chat model mid-conversation without starting over.
Multimodal input
Attach an image and ask about it; the assistant reads and responds.
Transparent credits
An always-visible balance and clear spend moments, starting with a free grant.
Fast creation shortcuts
Intent buttons above the input jump straight into image, audio or video.
Made to keep
Copy, share, regenerate and rate on every response for real day-to-day use.
Got something like this in mind?
We take products from a first sketch to the app store. Tell us what you're building and we'll tell you honestly how we'd approach it.