SenseAI — Hardware & Infrastructure Requirements

Pre-installation checklist for school IT administrators

SSQUARE AI SOLUTIONS PRIVATE LIMITED ssquareit.co.in
senseai-support@ssquareit.co.in

🖥️ Server / Dedicated Machine Requirements

Component	Minimum Requirement	Recommended	Notes
Processor (CPU)	Intel Core i7 (10th Gen+) or AMD Ryzen 7 5000+	Intel Core i9 / Xeon or AMD Ryzen 9	Minimum 8 cores, 16 threads for AI inference speed
RAM (Memory)	32 GB DDR4	64 GB DDR4/DDR5	AI model (llama3.1:8b) alone needs ~8 GB. With Live Quiz + 40 concurrent student sessions, 32 GB provides comfortable headroom. 64 GB recommended for schools with 100+ simultaneous users
Storage (OS + Software)	512 GB SSD (NVMe preferred)	1 TB NVMe SSD	SSD mandatory — HDD causes severe AI loading delays (5–10 min per query)
Storage (AI Models)	100 GB free space	200 GB free space	Software package: ~32 GB. AI models included in package. Document storage: varies by school usage
GPU (Optional)	Not required	NVIDIA RTX 3060 / RTX 5050 (8 GB VRAM+)	GPU reduces AI response time 6–10x. Measured performance: 5 questions in ~15–44 sec (GPU, warmed up) vs 11–16 min (CPU cold start). After first job, GPU stays fast. CPU-only mode works fully. NVIDIA GPU strongly recommended for school demos
Operating System	Windows 10/11 (64-bit, Pro recommended)	Windows 11 Pro / Windows Server 2019	WSL2 must be enabled. Pro edition recommended. Home edition works if WSL2 is enabled
Network Interface	1 Gbps Ethernet (LAN)	1 Gbps Ethernet + Wi-Fi backup	Server must have a static/reserved IP on school LAN (e.g., 192.168.1.100)

✅ Tested & Recommended Machine

✅ Lenovo LOQ 2025 — Tested & Verified by SSQUARE AI SOLUTIONS

Processor: AMD Ryzen 7-250 (AI Series) — 440 AI TOPS NPU
GPU: NVIDIA GeForce RTX 5050 — 8 GB VRAM (CUDA 13.0) — GPU acceleration confirmed working
RAM: 24 GB DDR5
Storage: 1 TB NVMe SSD
Display: 15.6" FHD 144Hz
OS: Windows 11 Home (WSL2 enabled)
Model: 83JG008NIN | Weight: 2.4 Kg
Performance (measured): 5 questions in ~15 sec, 10 questions in ~39 sec, 11 questions in ~44 sec (after GPU warmup)
Cold Start Note: First AQG job after system boot takes 11–16 minutes (GPU loading model into VRAM). All subsequent jobs are fast. Run one warmup job before any live demo or class session

⚠️ Critical: Do NOT use these configurations

Low-end or thin-and-light laptop as server — thermal throttling kills AI performance. High-performance gaming/workstation laptops (e.g., Lenovo LOQ, ASUS ROG) with dedicated GPU are acceptable
HDD (spinning disk) storage — AI model loading takes 5–10 minutes, unusable
Less than 32 GB RAM on CPU-only machines — AI models will crash with out-of-memory errors. Exception: machines with NVIDIA GPU (8 GB VRAM+) can run with 24 GB RAM as the model loads into GPU VRAM instead of system RAM
Windows Home Edition — Docker Desktop requires WSL2 to be enabled; Pro edition recommended for stability
Shared internet PC — AI inference is CPU-heavy; sharing with teacher workstation causes slowdowns

⚡ Performance Tips

Cold Start — GPU Model Warmup

After the server boots, the first AQG job takes 11–16 minutes because Ollama loads the AI model (~8 GB) into GPU VRAM for the first time. Once loaded, all subsequent jobs complete in 15–44 seconds. Always run one dummy AQG job before a live demo or class session to pre-warm the model. The model stays loaded as long as the system is running — no warmup needed between jobs.

🌐 Network Requirements

📡 LAN Setup (Required)

Switch with minimum 100 Mbps ports
Server must have a fixed / reserved IP
All teacher / student devices on same LAN
Router must allow local traffic (no inter-VLAN blocking)
Live Quiz: stable LAN essential — students answer in real-time; network drops cause missed responses
Router/switch must not block WebSocket connections (port 80 upgrade) — all standard routers support this by default

🔌 Power & UPS

Server must be on a dedicated UPS
UPS capacity: minimum 1 kVA for server alone
Power fluctuations cause database corruption
Sudden shutdown during AI generation = job failure
Recommended: APC / Luminous online UPS

💻 Computer Lab / Client Device Requirements

🖥️

Desktop / Lab PC

Any desktop or laptop with 4 GB RAM+. No software installation needed — students open a browser and access the system directly from the school server.

🔌

LAN / Network

All lab PCs must be connected to the same local network as the server (wired LAN recommended). No internet required — SenseAI runs fully offline on the school server.

🌐

Browser Requirement

Chrome 90+ or Firefox 88+ required. Edge Chromium supported. Internet Explorer not supported. No plugins or extensions needed.

✅ What You Do NOT Need

Internet connection — SenseAI is fully offline after installation
Cloud subscription or external server — everything runs on your school server
Client software on student/teacher PCs — just a web browser is sufficient
IT team for daily operation — designed for non-technical school staff

SenseAI — Pre-Installation Checklist

Complete this checklist before scheduling installation

SSQUARE AI SOLUTIONS PRIVATE LIMITED ssquareit.co.in

✅ Server Readiness Checklist

Please tick each item and share this completed checklist with SSQUARE AI SOLUTIONS before your installation date.

Dedicated server machine ready — not a shared workstation or laptop (i7, 32 GB RAM, 512 GB SSD minimum)
Windows 10/11 Pro installed — fresh installation preferred, OS updated to latest patches
UPS connected to server — minimum 1 kVA, server should survive 30-minute power cut
Static IP assigned to server — e.g., 192.168.1.100 (coordinate with your network admin)
100 GB free disk space confirmed — AI models, Docker images, and document storage
Server accessible from classroom PCs — ping test successful from at least one teacher PC
Administrator account credentials ready — needed during installation for software setup
Antivirus / Windows Defender exclusions set — exclude SenseAI installation folder from real-time scanning
USB 3.0 port available on server — installation is done via USB drive (minimum 64 GB, USB 3.0 required for speed)

🌐 Network Readiness Checklist

LAN switch with available ports — server connected via Ethernet (not Wi-Fi)
All classrooms reachable on same LAN — or at minimum teacher room + computer lab
Port 80 not blocked — this is the port used by SenseAI (standard web port, same as any website)
Firewall allows internal traffic — no firewall rule blocking 192.168.x.x to server IP
Wi-Fi available in classrooms/exam hall (optional) — any device (laptop, tablet, smartphone) connected to the same Wi-Fi can access SenseAI through a browser — no software installation needed on the device

👤 Administrative Readiness Checklist

List of teachers with names and email IDs — to create teacher accounts during setup
Subject structure decided — e.g., Physics, Chemistry, Biology, Maths for Class 11
Student list ready (optional) — roll number, name, class. Note: students can also join sessions using a PIN shared by the teacher — no individual student accounts required
Documents to upload identified — syllabus PDFs, NCERT chapters, school policies
IT person designated as Tenant Admin — one person responsible for ongoing user management

📋 Server Details to Share with SSQUARE AI SOLUTIONS

School Name	_______________________________________________
Server IP Address	___________________ (e.g., 192.168.1.100)
Processor	___________________	RAM	____________
OS Version	___________________	Storage Free	____________
TeamViewer ID	___________________	TeamViewer Pass	____________
Contact Person	___________________	Phone	____________
Preferred Install Date	_______________________________________________

📞 Ready to Install? Contact Us

Once you've completed this checklist, share it with us via email or WhatsApp. We'll schedule an installation slot (remote via TeamViewer or on-site for TN/nearby states) within 2–3 working days of receiving payment confirmation.

Email: senseai-support@ssquareit.co.in | Website: ssquareit.co.in