Make a sound and KAIKU paints what it heard. It listens through your microphone, draws the sound as it arrives, reads it back in plain words, and turns it into a picture anyone can come and browse.
KAIKU makes a picture from a sound. It listens through the microphone for up to twelve seconds and draws the sound on the screen while it listens. Then it measures four things: loudness, pitch, pace and brightness. It tells you in plain words what it heard, and AI paints a picture from that reading.
The audio stays in your browser. Only the four numbers and a short text about them leave the browser. KAIKU describes the sound. It never describes the person who made the sound.
All pictures go to a public gallery. Anyone who opens the app can browse the gallery without an account. Each picture shows the measured values and the words that KAIKU wrote about the sound.
Only the owner of the app can record and make new pictures. The owner signs in, starts a recording and makes a sound. Without AI, the app still listens, measures and describes the sound, but it does not paint a picture.
To make a picture, you need the account of the owner, a microphone and AI on that account. To browse the gallery, you need only a web browser. The app is in English and Finnish.
KAIKU is for people who are curious about how a sound could look, and for anyone who wants to browse pictures that sounds made.