Drag blocks to build real AI — it can see, hear, read, and recognise gestures — then wire it to real hardware (Arduino, ESP32) and make things that move. Free, no downloads.
Every robot in Maker is built from the same three moves. Kids snap them together like LEGO — no code.
A camera, a mic, or a sensor takes in the world — a face, a gesture, a sound, a word, a distance.
An AI brain turns what it sensed into a decision — "that's a smile", "the finger points up", "someone is close".
The decision drives real hardware — spin a motor, turn a servo, light an LED, sound a buzzer.
Drop a brain onto the canvas and it just works — most are pre-trained on millions of examples. Two you can even teach yourself.
Teach it your own classes from photos, then it recognises them live.
You train itSpots a face and reads expressions — smiling, surprised.
Pre-trainedTurns your fingertip into up / down / left / right + a pinch click.
Pre-trainedAlready knows ~80 everyday things — bottle, cup, person, phone.
Pre-trainedReads written words through the camera and reacts to them.
Pre-trainedSeparates you from the background — knows when someone's close.
Pre-trainedTeach it your sounds — a clap, a whistle — and it tells them apart.
You train itLearns body poses and gestures from your webcam.
You train itEvery model has a Try It Live window. Kids don't just use the AI — they look inside it.
Point the camera or mic and watch the AI guess in real time, with confidence bars for every class.
Peek at what the AI actually measures — edge filters, heat-maps, muscle-sliders, a fingertip's zones, the words it read.
A live little neural network shows the real features flowing to a decision — honest, animated, and true to how it works.
Connect an Arduino or ESP32 straight from the browser — no drivers, no IDE, no install. Then let your AI drive it.
Web Serial connects your board in one click. Upload the firmware once and you're set.
Drive DC motors, servos, and continuous-rotation wheels — build cars and arms.
Set pins, blink LEDs, sound buzzers — instant feedback for any decision.
Distance, motion, line, touch, sound, light, moisture — feed the real world back in.
Point your finger and the robot car drives that way — pinch to stop.
The Object Finder spots trash; a servo pops the lid open.
When it sees a smile, it waves and says hello.
Person Cutout senses someone close and flashes a warning light.
Mix a spoken word with a hand gesture to trigger different actions.
Hold up a card that says "GO" and the robot moves; "STOP" and it brakes.