OpenFocus: AI Study Agent

What it does#

OpenFocus is an AI study platform that brings scheduling, task management, and personalised tutoring into one workspace. Rather than a single assistant, it offers five AI agents with distinct personalities so the tutoring style can match how a student learns. Around the agents sit a smart calendar, goal-based progress tracking with a gamified achievement system, and AI reading tools that add inline annotations and generate quizzes.

Architecture#

The repository is a monorepo split into a complete frontend and an in-progress backend:

study-agent/
  frontend/   React 18 + TypeScript SPA
  backend/    Python service (in development)
  docs/

The frontend is the focus today and is fully deployed; the Python backend that will host agent orchestration is still being built out.

Frontend stack#

• React 18 with TypeScript for the UI.
• Vite as build tool and dev server (npm run dev serves on http://localhost:5173).
• React Router v7 for client-side routing.
• Zustand for lightweight, store-based state.
• Tailwind CSS v4 for styling.
• Motion for view and component transitions.

The Zustand-plus-React-Router combination reflects a client-first architecture: schedule, task, and conversation state lives in the browser and syncs to the backend API as it lands.

Status and deployment#

The codebase is roughly 70% TypeScript and 26% Python by line count, reflecting the mature frontend and the emerging Python backend. It is deployed on Railway at study-agent-production.up.railway.app, and grew its own component and state architecture from an initial Figma-exported template.

Why it matters#

It is an experiment in multi-persona tutoring: instead of one generic chatbot, the product models several teaching styles and lets the learner pick the one that fits, wrapped in the scheduling and tracking tools that keep a study habit alive.