Mansori

How Mansori teaches

Mansori is a free path of 50 short, interactive animations that take a child from noticing patterns at age 3 to training neural networks at 18. Here is how it works, what it covers, and how we keep children safe.

The approach

One idea per lesson, and no repeats. Every lesson teaches exactly one concept with its own animation, and each one names the earlier idea it builds on. A child who finishes the path has met the core ideas of computer science and machine learning in the order engineers would teach them.

Doing, not watching. Children tap, drag, wire circuits, train models and break things. The animation reacts instantly, which turns abstract ideas (a loop, a logic gate, a gradient) into something a child can see and feel.

Honest about AI. We show how AI actually works, including where it fails: a model that memorizes instead of learning, and a model trained on lopsided data that treats people unfairly.

No promises we can't keep. We don't claim to raise IQ or test scores. What the lessons give is steady practice with the habits engineers use: spotting patterns, breaking problems into steps, testing ideas, and fixing mistakes.

Built for children, private by design

Read the full privacy policy.

Using Mansori in a classroom

Each lesson takes 4 to 10 minutes and ends with discussion questions and a screen-free activity you can run with paper, cups or string. Lessons work on Chromebooks, iPads and phones with no installation and no student logins, which makes them easy to use in schools.

The full curriculum

#LessonAgesLearning goalBuilds on
1What Comes Next? Color Patterns3–5Spot a repeating pattern (red, blue, red, blue) and pick what comes next.—
2Big to Small: Putting Things in Order3–5Put objects in order from biggest to smallest, the first idea behind sorting.What Comes Next? Color Patterns
3Count the Lights3–5Count from 1 to 10 by turning on one light for each number.—
4Light On, Light Off3–5Learn that a switch can be only on or off, the tiny idea every computer is built from.Count the Lights
5Spot the Odd One Out3–5Find the one thing that is different from the rest.What Comes Next? Color Patterns
6Shape Sorting Bins3–5Sort circles, squares and triangles into the right bins, the start of classification.Spot the Odd One Out
7Robot Steps to the Apple3–5Give a robot a short list of arrow steps to reach the apple.Big to Small: Putting Things in Order
8If It Rains, Take an Umbrella3–5Match each kind of weather to what we need, learning simple if-this-then-that rules.Light On, Light Off
9Find the Way Through the Maze3–5Trace a path from start to finish and learn that some paths are dead ends.Robot Steps to the Apple
10Teach the Robot: Apple or Banana?3–5Show a robot examples until it learns to tell apples from bananas: the very first idea of AI.Shape Sorting Bins
11Algorithms Are Recipes6–8Put the steps of making a sandwich in order and learn what an algorithm is.Robot Steps to the Apple
12Loops Make a Dance6–8Use a repeat block so a dancer repeats moves without writing them again and again.Algorithms Are Recipes
13Find the Bug!6–8Run a robot's program, watch where it goes wrong, and fix the one bad step.Loops Make a Dance
14Counting With Light Bulbs: Binary6–8Use five light bulbs worth 16, 8, 4, 2 and 1 to make any number up to 31.Light On, Light Off
15Pictures Are Made of Pixels6–8Paint a picture square by square and see how screens show images with tiny pixels.Counting With Light Bulbs: Binary
16Sorting Animals by Swapping6–8Line animals up by height by only swapping neighbours, a real sorting algorithm.Big to Small: Putting Things in Order
17When Clicked: Events6–8Connect buttons to actions and learn that programs wait for events like a click.Loops Make a Dance
18Guess the Animal in Yes/No Questions6–8Follow yes/no questions down a tree to guess an animal, the way many computer decisions work.Shape Sorting Bins
19Secret Codes With a Cipher Wheel6–8Spin a cipher wheel to send and crack secret messages.Algorithms Are Recipes
20How AI Learns From Examples6–8Give a computer more and better examples and watch its guesses get more accurate.Teach the Robot: Apple or Banana?
21Variables Are Labeled Boxes9–11Store, change and read values in labeled boxes, and trace what a short program does.Algorithms Are Recipes
22If/Else: Run the Traffic Lights9–11Write if/else rules that run a traffic light safely, and see what happens when a rule is wrong.Variables Are Labeled Boxes
23Guess My Number in 7 Tries9–11Find a number from 1 to 100 in at most 7 guesses by cutting the range in half each time.Counting With Light Bulbs: Binary
24Logic Gates: AND, OR, NOT9–11Flip switches into AND, OR and NOT gates and predict when the lamp lights up.If/Else: Run the Traffic Lights
25Shortest Path on a Map9–11Watch a search spread out like a ripple to find the shortest route, the idea behind map apps.Find the Way Through the Maze
26Shrinking Pictures: Compression9–11Squeeze a picture into fewer numbers by counting runs of the same color.Pictures Are Made of Pixels
27How the Internet Sends Messages9–11Split a message into packets, watch them take different routes, and put them back in order.Shortest Path on a Map
28Letters Are Numbers Inside a Computer9–11Type a word and see each letter turn into a number and then into binary.Counting With Light Bulbs: Binary
29AI Guesses by Looking at Neighbours9–11Drop a new point on a chart and let its nearest neighbours vote on what it is.How AI Learns From Examples
30Coin Flips and Chance9–11Flip thousands of virtual coins and dice and watch randomness turn into a predictable shape.Count the Lights
31Functions Are Input-Output Machines12–14Feed numbers into mystery machines, guess the rule, and chain functions together.Variables Are Labeled Boxes
32Recursion Grows a Tree12–14Use a function that calls itself to grow a fractal tree, and see the base case stop it.Functions Are Input-Output Machines
33The Sorting Race: Why Speed Matters12–14Race bubble sort against merge sort and see why some algorithms scale and others crawl.Sorting Animals by Swapping
34Build an Adder From Logic Gates12–14Wire XOR and AND gates into a circuit that adds binary numbers, just like a real chip.Logic Gates: AND, OR, NOT
35Inside a CPU: Fetch, Decode, Execute12–14Step a tiny CPU through a program and watch memory, registers and the program counter change.Build an Adder From Logic Gates
36Hashing: How Passwords Stay Secret12–14See how a hash turns any text into a fingerprint, and why one tiny change scrambles it.Secret Codes With a Cipher Wheel
37The Game of Life: Simple Rules, Complex Worlds12–14Draw cells, press play, and watch four simple rules create gliders, oscillators and chaos.If/Else: Run the Traffic Lights
38A Perceptron Learns to Draw a Line12–14Watch the simplest artificial neuron adjust its weights until it separates two groups.AI Guesses by Looking at Neighbours
39Gradient Descent: Rolling Downhill to Learn12–14Tune the learning rate and watch a ball roll down an error curve to the best answer.A Perceptron Learns to Draw a Line
40How Image Filters Find Edges12–14Slide a 3x3 kernel over a picture to blur it, sharpen it, or find its edges.Pictures Are Made of Pixels
41Inside a Neural Network15–18Change inputs and weights and watch signals flow through layers of neurons to an output.A Perceptron Learns to Draw a Line
42Backpropagation: Training a Network Live15–18Train a small network on XOR in your browser and watch the decision boundary bend.Gradient Descent: Rolling Downhill to Learn
43Overfitting: When a Model Memorizes15–18Fit curves of rising complexity and see training error fall while test error climbs.Backpropagation: Training a Network Live
44Words as Points: Embeddings15–18Explore a map where similar words sit close together, and do word arithmetic like king - man + woman.Inside a Neural Network
45How Chatbots Pick the Next Word15–18Build a tiny language model from text and change the temperature to see creative versus safe choices.Coin Flips and Chance
46Attention: How Transformers Focus15–18Click a word and see which other words it pays attention to, the core trick inside ChatGPT and Claude.Words as Points: Embeddings
47K-Means: Finding Groups Without Labels15–18Drop centroids, step the algorithm, and watch unlabeled points organize into clusters.AI Guesses by Looking at Neighbours
48Reinforcement Learning: An Agent Learns a Maze15–18Reward an agent for reaching the goal and watch Q-values spread until it finds the best path.Shortest Path on a Map
49Public-Key Secrets With Paint Mixing15–18Agree on a shared secret in public using the paint-mixing picture of Diffie-Hellman.Hashing: How Passwords Stay Secret
50Biased Data, Biased AI15–18Train a model on lopsided data, see who it fails, and fix it by fixing the data.Overfitting: When a Model Memorizes

Contact

Questions, ideas for lessons, or found a bug? Email support@mansori.ai.