Nova by THOX.ai
The privacy-first edge AI computer for professionals. Run AI models locally with hardware-accelerated inference. For healthcare, legal, research, development, and beyond.
Founder pricing from $629 — reserve with a refundable $99.99 deposit
Views
front View
back View
side View
Technical Specifications
Enterprise-grade hardware engineered for any AI workload. Built for professionals who demand performance and privacy.
Form Factor
- Display
- 6.0-inch touchscreen, 720 × 1280
- Dimensions
- 100 × 160 × 23 mm
- Weight
- ~285g
- Chassis
- 6061-T6 aluminum, fanless
- Colors
- Matte Black · Space Gray · Arctic White
Compute
- Processor
- 8-core x86 compute
- Memory
- 16GB DDR5
- Storage
- 512GB NVMe (AES-256 encrypted)
- AI Acceleration
- Hardware-accelerated AI inference
AI Performance
- Processing
- 100% local, hardware-accelerated
- Latency
- Sub-50ms typical
- Context Window
- 256K–384K tokens
- Languages
- 32+ supported
- Pre-installed Models
- Vision, reasoning, OCR
Connectivity
- Wi-Fi
- Wi-Fi 6
- Bluetooth
- 5.2
- USB
- USB-C (data + display)
- Charging
- Qi2-compatible wireless charging
Power
- Battery
- 28Wh, 4–6 hour runtime
- Charging
- USB-C 45W Power Delivery
- Wireless
- Qi2-compatible
MagStack Clustering
- Interface
- Face-to-back magnetic alignment
- Device Discovery
- Automatic discovery
- Cluster Formation
- Automatic, no cables
- Max Cluster Size
- Up to 8 devices
- Combined Memory (8x)
- Up to 128GB
Compliance
- Architecture
- HIPAA, GDPR, ITAR-friendly
- Operation
- Works offline
- Data
- Stays on device
MagStack™ Clustering
Stack devices magnetically to combine RAM and compute power. Automatic discovery. Instant scaling. No cables, no configuration.
- Combined Memory
- 16GB
- Throughput
- 20-72 tok/s
- Max Model Size
- 32B
- Combined Memory
- 32GB
- Throughput
- 25-45 tok/s
- Max Model Size
- 70B
- Combined Memory
- 64GB
- Throughput
- 15-30 tok/s
- Max Model Size
- 100B+
- Combined Memory
- 128GB
- Throughput
- 10-20 tok/s
- Max Model Size
- 200B+
MagStack™ is a trademark of THOX.ai LLC.
Pre-installed Models
Ready to use out of the box. Hardware-accelerated inference for maximum performance.
THOX vision model
Vision, 32+ languages, edge AI
- Performance
- Fast inference
- Context
- 256K tokens
THOX reasoning model
Advanced reasoning, vision
- Performance
- Hardware-accelerated
- Context
- 128K tokens
THOX long-context model
Long context analysis
- Performance
- Hardware-accelerated
- Context
- 128K tokens
THOX OCR model
32-language OCR, code specialist
- Performance
- Hardware-accelerated
- Context
- 128K tokens
Compatible with thousands of open-source models. Hardware-accelerated inference delivers blazing-fast performance across all model sizes.
What's in the Box
Everything you need to get started.
- THOX.ai Nova device
- USB-C 45W Power Delivery charger
- USB-C to USB-C Cable (1m)
- Quick Start Guide
Powered by ThoxOS™
Purpose-built edge AI operating system optimized for maximum performance.
Hardware-Accelerated Inference
Optimized for maximum performance across all model sizes.
Secure by Design
Hardware root of trust, encrypted storage, verified boot chain.
Edge AI Compute Module
Purpose-built compute with native hardware optimization for local AI.
AI Runtime
- Hardware-accelerated inference
- Pre-installed models (8B/14B/30B/32B)
- Smart model routing
- OpenAI-compatible API
Ready for Any Workflow
- Intuitive web dashboard
- API access for any application
- CLI tools for power users
- Automatic updates (OTA)
ThoxOS™ is a trademark of THOX.ai LLC. All rights reserved.
Frequently Asked Questions
Got questions? We've got answers.
Three more SKUs. Same THOX experience, scaled up.
We’re currently wrapping up research and development on the THOX Edge Series™ and will be launching it right after our Nova flagship device ships. The line extends the THOX hardware family into workgroup, department, and enterprise-rack tiers — same ThoxOS, same MagStack clustering, scaled compute. Three SKUs, three chassis colors. Not part of the Founders campaign.
THOX Pro
32GB · 64 TOPS
Designed for small teams. Same drop-in API surface as Nova.
THOX Pro Max
64GB · 128 TOPS
Bigger memory ceiling for department-scale workloads.
THOX Pro Ultra
128GB · 256 TOPS
Rack-scale edge AI for regulated industries and 200B+ models.
