About
About YAKKI EDU
A small Israeli team building an AI-powered platform for learning English — with the conviction that pedagogy leads and technology serves.
Mission
English for every learner — on their terms
English opens universities, careers, and half the internet. Yet in mixed classrooms, learners are routinely held back by things that have nothing to do with their ability: instructions in a language they don’t yet read, content about topics they don’t care about, and one-size-fits-all exercises pitched at nobody in particular.
YAKKI’s answer is generation instead of libraries: exercises produced on demand for this learner, this topic, this level — with the interface in the learner’s first language, and adaptations for learners with dyslexia.
How we work
Principles
Pedagogy first
Every game format starts from a documented learning mechanism — retrieval, repetition, feedback — and only then becomes software. The science behind the platform is published on the Methodology page.
Honest status labels
Everything on this site is labeled: Available Now, Pilot Ready, or In Development. We would rather undersell than mislead a school.
Tested relentlessly
Every game passes systematic play-testing — played to completion, with deliberate mistakes, on real devices — before it is offered to a classroom.
Inclusion as architecture
First-language interface and dyslexia adaptations are core platform features, not bolted-on extras.
Where we are
Stage: working product, pilot phase
The platform is real and demonstrable today: two products (Polygon and GuestDay), 100+ game formats, five interface languages. We are now building partnerships with schools, researchers, and institutions in Israel for pilots and validation studies. The full roadmap — including three product assemblies and an individual pedagogical model — is public.
Scientific grounding
Standing on published research
The platform’s design draws on international research in second-language acquisition and on work by Israeli researchers in EFL teaching, reading acquisition in Hebrew and Arabic, dyslexia, and learning analytics. We cite this work transparently and do not claim endorsement by its authors. We actively seek research collaborations to evaluate the platform’s effectiveness.