Artificial intelligence for science education.

Prepare a lesson explaining the derivative as the slope of the tangent line, for a final-year high school class
Mathematics

The derivative is a slope

Project co-funded by the European Union

PrismaX is our AI agent: it builds interactive courses for teachers and tailored lessons for learners, with animated figures and self-checking exercises.

Free to start.

Backed byGoogle for Startups

Ready-made lessons to start from.

Mathematics, physics, statistics, data analysis, machine learning.

The derivative is a slope
MathematicsThe derivative is a slope
Gradient descent: how a model learns
Machine learningGradient descent: how a model learns
Electric field: seeing the invisible
PhysicsElectric field: seeing the invisible
The Galton board: from chance to the bell curve
ProbabilityThe Galton board: from chance to the bell curve
Reinforcement learning: learning by trial and error
Artificial intelligenceReinforcement learning: learning by trial and error
Sine and cosine: from the circle to the wave
MathematicsSine and cosine: from the circle to the wave
A/B tests: how many users you need
StatisticsA/B tests: how many users you need
Physics in code: an orbit in a few lines
ProgrammingPhysics in code: an orbit in a few lines
Quadratic functions: vertex, roots and the best price
MathematicsQuadratic functions: vertex, roots and the best price
Exponential growth and logarithms
MathematicsExponential growth and logarithms
Parabolas and quadratic inequalities
MathematicsParabolas and quadratic inequalities

Whole courses, chapter by chapter.

Every chapter has an explanation, an animation and exercises.

How a machine learns

The line that follows the data

For a machine, learning means finding the numbers that make it wrong as little as possible. Here there is only one: the slope ww of the line y=wxy = wx, which must pass as close as possible to six points.

For each point we take the residual, how far the point lies above or below the line. The error is the mean of the squared residuals:

L(w)=1n∑i=1n(yi−w xi)2L(w) = \frac{1}{n}\sum_{i=1}^{n} \left(y_i - w\,x_i\right)^2

where (xi, yi)(x_i,\,y_i) are the points and nn is how many there are.

Drag the line or the point on the curve: the error changes with w.

Try it

Exercise

With the line y=xy = x and the two points (1, 1)(1,\,1) and (2, 3)(2,\,3), what is the error LL?

0.5Check

Correct!

The residuals

1−1=03−2=11 - 1 = 0 \qquad 3 - 2 = 1

The mean of the squares

L=02+122=0.5L = \tfrac{0^2 + 1^2}{2} = 0.5

Problems and solutions, step by step.

New variants whenever you want, each one checked automatically.

Exercise

Solve 2(x−3)=x+42(x - 3) = x + 4.

AnswerCheck

See how each student answers.

Teacher only

Students’ results

sample data

ChapterThe derivative4 students · 26 answers checked

By student

StudentRightWrong
Emma R.51
Jack B.54
Sophie L.40
Liam T.43

Recent activity

10:37
Sophie L.Right(6)
Check · Ex. 1
10:33
Liam T.—
Opened the solution · Ex. 2
10:31
Liam T.Wrong(12)
Check · Ex. 2
Try PrismaX

Free to start, no credit card.

How we built PrismaX.

A first call to the model plans the request in JSON; the next ones write the lesson, the figure and the exercises. The code runs in an isolated sandbox and, if it fails, goes back to the model with the error.

Request
Plan
Lesson
Figure
Exercises
Sandbox
Chapter

Development

Claude CodeNext.jsTypeScript

Infrastructure

CloudflareSupabaseStripe

Models

OpenRouterGeminiOpenAI

Figures and math

three.jsJSXGraphKaTeXmath.js

Our services.

For schools, universities, companies and startups.

StudentRightWrong
Emma R.51
Jack B.54
Sophie L.40

Course licenses

PrismaX for everyone enrolled and every teacher in the course, with restricted access and results by student and by exercise. For companies too, with seats for in-house trainers.

We start with a pilot on one class.

Turnkey interactive module

We build it for you: animations, exercises with automatic checking, an entry and a final test, with your brand.

What do you want to change or add?

Workshop

Creating interactive lessons with AI: hands-on training for teachers and trainers, with one year of access included.

More courses and sites? Contracts with your brand only and a dedicated contact person.

AI agents

We also build agents, data analysis and products on request, with processes and technologies similar to those we used to build PrismaX.

About us.

Edukai was founded in Rome in November 2023 to develop intelligent, accessible technologies that help startups, universities, companies and young talent design solutions with high technological, educational and social impact.

Andrea De Cosmo

Andrea De Cosmo

Founder and CEO, Edukai

Instructor at Temple University Rome

Contact us.

So we can point you to the right contact: you are writing as…