Cookie preferences

HundrED uses cookies to enhance user experiences, to personalise content, and analyse our web traffic. By clicking "Accept all" you agree to the use of all cookies, including marketing cookies that may help us deliver personalised marketing content to users. By selecting "Accept necessary" only essential cookies, such as those needed for basic functionality and internal analytics, will be enabled.
For more details, please review our Cookie Policy.
Accept all
Accept necessary
keyboard_backspace Back to HundrED
Lasse Leponiemi

Chairman, The HundrED Foundation
first.last@hundred.org

Calcularis

Neuroscience-based adaptive maths that builds number sense in children with dyscalculia

About five percent of children have dyscalculia, and many more leave early primary without secure number sense, yet a teacher with 25 children cannot re-sequence number ranges for each one, so the usual answer is more of the drill that has already failed. Constructor Calcularis is an adaptive maths programme for children aged 6–10, developed at ETH Zurich and evaluated with neuropsychologists at.

Overview

Information on this page is provided by the innovator and has not been evaluated by HundrED.

Updated August 2026
Web presence

2021

Established

4

Countries
Students early
Target group
That mathematical difficulty stops being read as a character flaw. Dyscalculia is roughly as common as dyslexia and far less recognised, so a child who cannot hold a number line in their head is told they are not trying and believes it. We would like the default response to a struggling child to be diagnostic rather than repetitive: establish which representation of number has not formed, and train that, at the pace that child needs, before the child concludes they are simply bad at maths. More broadly, we would like evidence to be the price of entry in educational technology. Calcularis exists because a research group spent years measuring whether it worked and published the results, including the inconvenient ones. If any school could ask any provider for a randomised controlled trial and expect an answer, children would get better tools and teachers would waste less time on ones that do not work.

About the innovation

Why did you create this innovation?

The work began with one child. In 2002 Markus Gross, professor of computer science at ETH Zurich, started looking for a better way to help his own son, who has dyslexia. The research group that grew out of that question — later the ETH spin-off Dybuster AG, found the same principle applied to mathematics.

Roughly five percent of people have dyscalculia, a neurological difficulty in processing numerical information, and many more children never build secure number sense in early primary. Unlike dyslexia, dyscalculia is barely recognised, so it is identified late or misread as low effort — and the standard response is more practice at the level the child has already failed. What these learners need is close to the opposite: individually paced, multi-sensory repetition that rebuilds number itself.

An ordinary classroom cannot deliver it. A teacher of 25 children cannot re-sequence the progression of number ranges per child, or diagnose which mental representation of number has not yet formed. Parents who want to help have no reliable way to see where their child is stuck, or whether practice is working.

So we set out to take what neuroscience knows about how number sense develops and turn it into 20 minutes a day a six-year-old will actually do, with an adult able to see what is happening. Weak mathematics does not stay in the classroom: it narrows occupational routes and it shapes how children see their own intelligence. That is what we are trying to change.

What does your innovation look like in practice?

A child works in short sessions, around 20 minutes, several times a week. The published protocols used at least three to four 20-minute sessions weekly, and that is the dose we recommend.

Each session picks up where the intelligent tutoring system left off. The system holds a model of what this child can and cannot yet do; after every exercise it updates that model and chooses the next task, moving number range and exercise type up or down. A child who is not secure with quantities below 20 does not get hundreds; a child who has that automated is not held there. No adult has to set a level.

Number is presented in several representations rather than as arithmetic to be memorised: as a quantity, a spoken word, a written numeral and a position on a number line, with games that require moving between them. That is what multi-sensory means in practice, and it follows the four-step model of number development (von Aster & Shalev, 2007). The exercises are game-framed, because the volume of repetition has to be tolerable for a six-year-old.

Adults work in the Coach dashboard: per-child progress, where the errors cluster, time on task, plus printable reports and certificates. In a school, a teacher typically assigns Calcularis as the individual practice slot inside the maths lesson or as targeted support; at home, a parent makes it a daily routine. Reading the dashboard needs no specialist training, which is deliberate — the adults using it usually are not specialists.

How has it been spreading?

Calcularis reached families and schools first in Switzerland, through the ETH Zurich spin-off Dybuster AG (2007), and spread largely through the schools, therapists and clinicians who had seen the University of Zurich and Children's Hospital Zurich studies. Evidence, rather than marketing, has been the main channel throughout — the randomised controlled trial published in Frontiers in Psychology in 2020 is still the reason most institutional conversations start.

By the time Constructor acquired Dybuster in October 2021, close to 200,000 children had used the two applications — Calcularis and the spelling programme now sold as Grafari — between them. The programme is available in English, German, French, Italian, Spanish and Dutch, and in use beyond Switzerland in Germany, the United States, Canada, Brazil and India.

Since 2021 it has been distributed as part of Constructor Tech, which reports 2,000+ partners and customers and 2.5M+ users across its platform. That gave Calcularis an institutional school channel it did not have as a standalone Swiss product, alongside the direct parent subscriptions it has always had.

How have you modified or added to your innovation?

Calcularis reached families and schools first in Switzerland, through the ETH Zurich spin-off Dybuster AG (2007), and spread largely through the schools, therapists and clinicians who had seen the University of Zurich and Children's Hospital Zurich studies. Evidence, rather than marketing, has been the main channel throughout, the randomised controlled trial published in Frontiers in Psychology in 2020 is still the reason most institutional conversations start.

By the time Constructor acquired Dybuster in October 2021, close to 200,000 children had used the two applications, Calcularis and the spelling programme now sold as Grafari, between them. The programme is available in English, German, French, Italian, Spanish and Dutch, and in use beyond Switzerland in Germany, the United States, Canada, Brazil and India.

Since 2021 it has been distributed as part of Constructor Tech, which reports 2,000+ partners and customers and 2.5M+ users across its platform. That gave Calcularis an institutional school channel it did not have as a standalone Swiss product, alongside the direct parent subscriptions it has always had.

If I want to try it, what should I do?

Start with the product page: constructor.tech/products/learning/calcularis: age range, the pedagogy, and what the Coach dashboard shows.

If you are a parent: create an account, add a child profile, and let your child run the first sessions unaided. The system needs a few sessions to place the child correctly, so resist intervening when early tasks look too easy — that is the placement working. Open Coach after the first week and look at where the errors cluster rather than at the score.

If you are a school: start with one class or one support group, not a whole-school rollout. Choose the children for whom ordinary maths practice is not working. Give them a fixed 20-minute slot at least three times a week for a full term this is the dose used in the published trials, and sporadic use will not reproduce their results. Have the teacher review Coach weekly, and at the end of the term compare those children's progress against your own records for equivalent children in previous years before deciding whether to widen it.

To talk to us: get in touch through constructor.tech and ask for schools onboarding, which covers account provisioning for classes and a walkthrough of the dashboard.

Implementation steps

Set up the child's profile
Create an account and add a profile for each child, with their age and school year. Nothing else needs configuring: no level is set by an adult, because the intelligent tutoring system establishes where the child is during the first sessions. If you are a school, ask for schools onboarding and provision accounts for the class or support group in one go rather than individually.
Let the first sessions place the child
Have the child work through the first few sessions unaided, roughly 20 minutes each. Early tasks may look too easy that is the system establishing which representations of number are secure, and intervening or skipping ahead corrupts the placement. Do not sit with the child correcting them during this stage.
Have the child work through the first few sessions unaided, roughly 20 minutes each. Early tasks may
The published trials used at least 42 sessions of 20 minutes over a maximum of 13 weeks, so this is the dose the results are based on. In a school, book it as a recurring slot inside the maths lesson or the support period; at home, attach it to an existing daily routine. Sporadic use will not reproduce the published gains, and is the most common reason a trial disappoints.
Read the Coach dashboard weekly, looking at error patterns not scores
Open Coach once a week and look at where the errors cluster and how long the child actually worked, rather than at a score. The clusters tell a teacher or parent which representation of number is not yet formed which is the information a classroom cannot otherwise produce. Print a report or certificate for the child when something is achieved; the motivational effect is part of the design.
Review after a full term against your own records
At the end of the term, compare these children's progress against your own records for equivalent children in previous years, then decide whether to widen it. For a school this is the decision point for moving from one group to several; for a parent it is where you decide whether the routine is worth keeping.

Spread of the innovation

loading map...