IoT_Master

Raspberry Pi IoT & AI Teaching Platform

A modular, reusable, portable terminal for classroom IoT and AI demonstrations, built on a Raspberry Pi 4B with the official 5” touchscreen.

Quick start (for adults)

Students should follow the walkthrough on the Pi IoT Academy website. For anyone who just wants the commands:

# On a fresh Pi with Raspberry Pi OS 64-bit:
sudo mkdir -p /opt/iot-platform && sudo chown $USER:$USER /opt/iot-platform
cd /opt/iot-platform
git clone https://github.com/Computing-Hub/IoT_Master.git .
./scripts/bootstrap-pi.sh        # installs Docker, adds user to docker group
# (log out + back in for the group change)

cp .env.example .env
nano .env                         # set MQTT_PASSWORD; ANTHROPIC_API_KEY can wait
./scripts/bootstrap-mqtt-passwd.sh

docker compose up -d
docker compose ps

When all five containers report healthy:

URL	What
http://<pi>/	Vue placeholder dashboard
http://<pi>:1880/	Node-RED
http://<pi>:8000/health	API health check
http://<pi>:8001/health	AI service health check
mqtt://<pi>:1883 (auth required)	Mosquitto MQTT broker

Goals

• A single reusable platform, not a collection of one-off projects.
• Stable, portable, lightweight, and classroom-friendly.
• Modular: future modules (sensor hub, weather station, greenhouse monitor, bike dashboard, AI assistant) all reuse the same MQTT bus, FastAPI backend, SQLite store, and Vue frontend.
• Realistic for Raspberry Pi 4B hardware — no exotic peripherals in v1.

Non-goals (v1)

• No SIM, LTE, SMS, or smartphone integration.
• No flexible/foldable displays or custom display drivers.
• No local AI inference — cloud APIs only (Claude as reference provider).
• No multi-arch container builds — arm64 only.

Documents

Document	Purpose
docs/architecture.md	System architecture and Architecture Decision Records (ADRs).
docs/hardware-runbook.md	Step-by-step Phase 1 hardware setup (SD card, OS, touchscreen, networking, power).
docs/folder-layout.md	Target repository folder structure for code, modules, and infrastructure.
docs/mqtt-conventions.md	MQTT topic naming, payload schema, QoS, and retention rules.
docs/database-design.md	SQLite schema, repository abstraction, and migration path to Postgres / TimescaleDB / InfluxDB.
docs/ai-integration.md	Claude Haiku integration: user-triggered “Explain” button, prompt template, and persistence.
docs/implementation-plan.md	Phased delivery plan from hardware bring-up through to the first classroom module.

Status

Phase	Title	Status
1	Core device (hardware, OS, touchscreen)	Runbook ready
2	Core services (Docker, MQTT, Node-RED, FastAPI, Vue)	Built
3	MQTT core	Built
4	Database core	Built
5	Dashboard core	Built
6	AI API integration (Claude Haiku, “Explain” button)	Built
7	First module — Classroom Sensor Hub	Built

Working language

British English throughout (e.g. colour, organise, behaviour).