Skip to content
🔒 Privacy-First AI

AI that respects solicitor-client privilege.

Draft briefs, analyze contracts, and research case law — without exposing a single word to the cloud.

Contract analysis with zero data exposure Case law research on-device Solicitor-client privilege by architecture

Launching soon · Apple Silicon optimized · Canadian built

[ Nodus app interface — screenshot placeholder ]

Capabilities

Everything you need,
nothing you don't.

Six core capabilities designed for professionals who can't compromise on confidentiality.

🔒

Private Chat

Converse with AI that never sees your data. Full inference runs on your hardware — nothing leaves your machine.

📄

Document Intelligence

Ask questions about contracts, medical records, financial statements. RAG-powered understanding with zero cloud exposure.

🔍

Smart Search

Semantic search across your document library. Find what you need by meaning, not just keywords.

📋

Profession Templates

Pre-built workflows for legal briefs, clinical notes, financial analyses, and psychological assessments.

☁️

Cloud Escalation

When local models aren't enough, escalate to frontier models. Your data is scrubbed of PII before it ever leaves.

💻

Multi-Platform

Native macOS with Apple Silicon optimization. Windows coming soon. Your data stays on your device.

Privacy Architecture

Six layers between
your data and the cloud.

Privacy isn't a policy — it's architecture. Every query passes through six protection layers before anything reaches the network.

Your Device

🔒

Local Inference

AI runs entirely on your hardware. No network calls.

📄

Document RAG

Your documents indexed locally. Never uploaded.

PII Scrubbing Pipeline

🔍

PII Detection

GLiNER NER model identifies sensitive entities.

🔄

PII Replacement

Names, addresses, IDs replaced with synthetic tokens.

📋

Audit Logging

Every scrubbing action logged locally for compliance.

Cloud (Scrubbed Only)

☁️

Cloud Relay

Scrubbed query sent to frontier model via Toronto relay.

↑ Your device never shares raw data ↑

Why Ferox

Built by someone who
understands the stakes.

Founder

🎖️

Military Veteran

Canadian Armed Forces — security-cleared operations

🚑

Primary Care Paramedic

Frontline healthcare — understanding regulated environments

Model Provenance

Gemma 3 12B

Google — Local inference

Claude (Anthropic)

Anthropic — Cloud escalation

GLiNER

Open source — PII detection

Compliance

PIPEDA (Federal)
PIPA (Alberta)
Law 25 (Québec)
Zero Telemetry
100% Local Processing
Canadian Data Sovereignty
Pricing

Simple, transparent
cloud pricing.

Every plan includes full local AI. Cloud tiers add frontier models through our PII scrubbing pipeline. Pricing announced at launch.

Professional

Coming Soon

Full AI for everyday professional work.

Cloud Credits

Included monthly allocation

  • Full local AI inference (Gemma 3 12B)
  • Monthly cloud credit allocation
  • Standard cloud models
  • PII scrubbing pipeline
  • Hardware-backed encryption
  • Full audit logging
Get Early Access
Most Popular

Business

Coming Soon

Maximum capability for demanding work.

Cloud Credits

Expanded monthly allocation

  • Everything in Professional
  • Expanded cloud credit allocation
  • All models including Claude Opus
  • Priority cloud processing
  • Advanced document analysis
  • Early access to new features
Get Early Access

Enterprise

Custom

Tailored for teams and organizations.

Cloud Credits

Custom allocation

  • Everything in Business
  • Custom credit allocation
  • Priority queue processing
  • Dedicated support
  • Custom deployment options
  • Volume licensing
Contact Us

Advanced models like Opus use more credits per query. Credits reset monthly. Pricing announced at launch.

Need more credits?

Top-up packs available at launch. Credits never expire.

Coming Soon

Get started in
under two minutes.

Ferox Nodus is launching soon. Join the early access list to be first in line.

No spam. Unsubscribe anytime.

1

Install

Download and install on macOS or Windows.

2

Download Model

Gemma 3 12B downloads on first launch (~7 GB).

3

First Query

Start chatting — everything runs locally.

System Requirements

macOS Apple Silicon (M1+), 14.0 Sonoma or later
Windows x86_64 with AVX2, Windows 10/11 (NVIDIA GPU recommended)
RAM 16 GB minimum (24 GB recommended)
Storage ~8 GB for models + app