G3N is an AI tutor built around the syllabus a student actually sits: Ghana's GES, BECE, WASSCE and CTVET, Nigeria's NERDC, Cambridge IGCSE and A Level, and IB. It runs two ways: in the cloud for full power (stronger models, image generation, premium voice), and on-device when the connection isn't there, so lessons, quizzes and marking keep working offline. Plus document summaries, live caption and translate, a reader mode for ages 5-8, and Braille- and dyslexia-friendly output.
Hi Product Hunt, I'm George, building G3N from Tema, Ghana.
Most AI tutors are general-purpose chat with a school skin on top. G3N is built around the syllabus a student actually sits: Ghana's GES, BECE, WASSCE and CTVET, Nigeria's NERDC, Cambridge IGCSE and A Level, and IB, down to real past papers.
It runs two ways, and that turned out to matter more than I expected. Online, students get the full cloud stack: stronger models, image generation, premium voice. Offline, the on-device model keeps lessons, quizzes and marking working when the data runs out mid-lesson. Same app, same syllabus, different engine underneath.
The part I did not expect to matter as much as it does is accessibility. Once the tutor could speak and listen, blind and low-vision students could use the whole app hands-free: voice quizzes, a camera that describes what it sees, Braille-friendly output, dyslexia-friendly text. That became core, not a setting.
It's running in technical institutes in Ghana today, and the same build serves international schools on IB and Cambridge. I'd love your feedback.
Hi Product Hunt, I'm George, building G3N from Tema, Ghana.
Most AI tutors are general-purpose chat with a school skin on top. G3N is built around the syllabus a student actually sits: Ghana's GES, BECE, WASSCE and CTVET, Nigeria's NERDC, Cambridge IGCSE and A Level, and IB, down to real past papers.
It runs two ways, and that turned out to matter more than I expected. Online, students get the full cloud stack: stronger models, image generation, premium voice. Offline, the on-device model keeps lessons, quizzes and marking working when the data runs out mid-lesson. Same app, same syllabus, different engine underneath.
The part I did not expect to matter as much as it does is accessibility. Once the tutor could speak and listen, blind and low-vision students could use the whole app hands-free: voice quizzes, a camera that describes what it sees, Braille-friendly output, dyslexia-friendly text. That became core, not a setting.
It's running in technical institutes in Ghana today, and the same build serves international schools on IB and Cambridge. I'd love your feedback.