What it is
CognitiveFlow gauges how focused you are in real time by reading the rhythm of your typing and input — no camera, no surveys.
It learns the difference between being in the zone and starting to struggle, so an app can respond: ease off and offer help when you're stuck, and stay out of the way when you're locked in.
It grew out of my undergraduate research in affective computing and human-computer interaction — quantifying cognitive load from the input patterns a person already produces.
How it works
Signals you already produce
Focus is inferred from typing rhythm and input patterns — no webcam, no self-report surveys.
Real time
The model distinguishes flow from struggle as it happens, rather than reporting on a session after the fact.
Built to act on
Ships as a VS Code extension built on the editor's API, so an editor can offer help at the right moment and stay quiet at the wrong one.
A scoring engine behind it
A FastAPI backend takes live keystroke and cursor data and turns it into the focus metrics the extension reacts to.
Built with
TypeScript · VS Code API · Python · FastAPI
· Developer tooling / machine learning — built by Musa Alanssari, a software engineer in Houston, Texas. Jan 2026 — Present.
Next project