Amaka AI · amaka.app · A DaveLabs product
Making the abstract tangible.
Language models were step one. Amaka AI is a visual reasoning engine — it draws, animates and explains any complex topic the way a good teacher does at a whiteboard, in fourteen languages, and it can be interrupted mid-explanation.
The problem
A textbook diagram is correct and static. A language model answers a physics question with three accurate paragraphs. Neither builds intuition.
Millions of students fail every year not because they lack intelligence, but because every tool they have gets the facts right and the teaching wrong. Understanding a mechanism means being able to picture it moving — which is what a good teacher builds at a whiteboard, one stroke at a time.
Languages, including Wolof, Mandinka, Fula and Jola
Launch-phase places, open now at amaka.app
Primitives available to a single explanation
What it does
It reasons in pictures.
It draws while it explains
Equations, chemical reactions and physics mechanics become animations generated in real time — planned per question, not retrieved from a library.
It can be interrupted
Ask a question halfway through and the board answers, then carries on. A recorded video cannot do this. A class of forty rarely can either.
It speaks your language
Fourteen languages including Wolof, Mandinka, Fula, Jola, Hausa, Igbo, Yoruba and Swahili — the same lesson, spoken differently.
It adapts to the learner
The same concept explained one way for a twelve-year-old and another for a second-year undergraduate.
It stays consistent
A fixed visual vocabulary means a hundred explanations look like one coherent teacher, not a hundred different styles.
It works on the device you own
No specialist hardware, no install. Open amaka.app and ask.
How it works
The Grammar of Graphics pipeline.
01
Decompose the concept
The question is broken into the smallest set of ideas that must land in sequence. Each becomes a section with its own visual goal.
02
Choose the visual grammar
Every section maps to primitives — an axis, a vector, a body, a label — drawn from a deliberately small vocabulary.
03
Score the timing
Strokes are timed against the narration, so a label arrives as the word is spoken rather than three seconds early.
04
Render and stay live
The board renders progressively and holds its state, so a follow-up attaches to what is already on screen.
DaveLabs Horizon Summer '26
See it work.
Unveiling Amaka AI | Official Presentation & Live Demo (DaveLabs Horizon Summer ’26)
Amaka AI Live Demo | The Visual Reasoning Engine for Education
“We are not just building tools; we are realizing a dream of technological self-reliance and unleashing human potential on a global scale.”
Godswill Iyke Dave, Founder and CEO, DaveLabs
For schools and institutions
The teacher a student can keep asking after the bell.
Amaka AI is rolling out to schools, ministries and educational institutions across The Gambia and West Africa. It does not grade, rank or profile students, and it does not replace a teacher’s judgement.
- Access for every student in the cohort, on the devices they already have
- Coverage across mathematics, physics, chemistry and biology
- Explanations in the languages the cohort actually speaks
- A structured onboarding session for teaching staff
- No student profiling, ranking or scoring
Amaka AI is live. The first 1,000 learners get in now.
Launch-phase access is capped while we scale server capacity. Schools and institutions can arrange a structured rollout with our team.