Project case study

Mentis

An experimental coding environment for AI-assisted learning.

Next.js · Monaco · Python · Go
On this page

Overview

Mentis combines an interactive coding environment and AI-assisted instruction. The supplied prototype stack includes Next.js, Tailwind, Monaco, a Python AI service, and a Go API. Earlier prototyping used Gemini 2.5 Flash Lite.

Architecture and work

The browser owns the editor; the application API and AI service provide the instruction path. Keeping their responsibilities distinct makes behavior easier to inspect.

Current state

Experimental. No production scale, measured learning-outcome result, or deployed integration is claimed.

Challenges and next steps

Editor state, latency, feedback quality, and execution behavior must cooperate. Documenting actual flows, evaluating instruction quality, and publishing examples remain useful next steps.

Artifacts

Verified source, demo, and editor screenshot links have not been supplied.

Last updated 01 Oct 2026Discuss this work