Pole Street AI is building machine-learning models that read early signs of mood disorders from low-cost wearables, so that detection and crisis prevention are affordable to Africans.
On June 10, 2026, we embarked on a quest to build tech for detecting bipolar disorder symptoms before they are evident. Along the way we learned about the shape of the mental health space and the solutions that could help.
There is a global mental health crisis. It is even worse in Africa, where there aren’t enough mental health workers for the continent, and the cost of healthcare is forbidding for most of the population.
We know the problem. The problem is glaring and over-specified. We are working on the solution: accessible mental healthcare detection and tracking solutions for everyone on the continent.
We are not the first to realize this: the internet is filled with research into digital phenotyping: the science of passively collecting behavioural data using personal digital devices.
We believe technology is finally at a stage where it might be possible to create reliable tools for tracking the mental health and moods of real humans outside of a laboratory.
We are embarking on a journey to transform all that research into a working product for individuals, caregivers, psychiatrists and healthcare providers.
We are doing this by applying artificial intelligence to high-correlational mental health signals (heart rate variability, sleep, steps, sunlight exposure and speech analysis) read off wearables and smartphones.
We built a model for detecting mood signals off the COLMI R02 smart ring using open-source technology. We observed the strengths and limitations of the ring form factor.
We are now onboarding volunteers to test our mood detection models using their smartphones and wearables.