Introduction

About Slothub (Slothub)

Slothub (Slothub) is an intelligent online learning and management platform designed to unify the teaching and learning experience. It replaces disconnected tools (gradebooks, chat groups, assignment forms) with an all-in-one ecosystem enhanced by Artificial Intelligence.

The platform leverages powerful technologies including:

  • AWS Bedrock AgentCore - Powering Slozy Assistant, a conversational AI tutor that uses LangGraph ReAct Agent architecture and RAG (Retrieval-Augmented Generation) against a PostgreSQL pgvector database to ensure accurate, curriculum-based answers.
  • OpenAI API - For automated document parsing and generative content (exercises, learning roadmaps).
  • Amazon S3 - For secure, structured storage of uploaded course materials (PDF documents) and generated assets.
  • Java Spring Boot - Providing a robust REST API backend for core educational domain logic (classrooms, users, assignments, timetables).

Workshop Overview

In this workshop, you will learn how to initialize, configure, and locally deploy Slothub’s microservice architecture:

  • Frontend (frontend/): React 18, TypeScript, Vite, Tailwind CSS. This component handles the user interface and dynamically renders AI output widgets.
  • Backend (backend/): Spring Boot 3.5, Java 17. Handles user authentication (JWT/OAuth2), data management, and the core educational domains.
  • AI Service (AI/): Python 3.12, FastAPI. Responsible for automated content generation, S3 file uploading, and proxying requests to the AI model.
  • Agent-Core (agent-core/): Python 3.10, LangGraph. The brain behind the Slozy AI Assistant, running locally with MemorySaver for testing or AWS Bedrock for production.
  • Database: PostgreSQL with pgvector extension running easily via Docker Compose.

Architecture Overview

The Slothub platform follows a microservice/Agentic AI approach:

  1. User Interaction: Users access the React frontend.
  2. Core Logic: Frontend calls the Spring Boot API for authentication and classroom data.
  3. AI Workloads: Frontend/Backend delegates content generation requests to the FastAPI Python service which calls OpenAI APIs and interact with AWS S3.
  4. Intelligent Tutoring: The chat component interacts with the Agent-Core (Slozy) to perform RAG over PostgreSQL pgvector context.
  5. Data Layer: All services share a unified PostgreSQL database managed via Docker Compose.

Slothub Architecture