Turnkey Appliance & Support
Simple pricing. Dedicated hardware. Zero cloud token bills.
One transparent upfront cost for the physical 64GB hardware, plus an optional annual assurance contract for bi-monthly updates and model upgrades.
The Physical Box
SAFi Appliance
$4,500
One-time purchase. Fits under standard $5,000 corporate card limits.
- Hardware: 64GB unified memory, 1TB NVMe Gen4 SSD, sub-120W peak power, 1-year replacement warranty
- Pre-loaded models: Qwen 27B, Qwen 35B MoE, Gemma4, ready on first boot
- The IT Triad: Sysadmin, Security Engineer, and Software Engineer agents
- 100% Local & Private: Zero cloud API keys, zero external data egress, zero token billing
- Enterprise network ready: Hardened Debian OS, pre-configured local HTTPS, browser console
Continuous Assurance
Annual Assurance
$1,440 / year ($120/mo)
Software maintenance, model refreshes, and direct engineering support.
- Bi-monthly stable upgrades: Verified software and policy releases every two months including TCB integrity verifications
- Hardware-matched model updates: Upgrades when newer, more capable open-weight models emerge suitable for the 64GB hardware envelope
- Direct email support: Responsive support from our engineering team for agent workflows and system-related issues
- Automated backup utilities: Disaster recovery tools and audit ledger integrity verification routines
Direct Ordering & Verification
Request your SAFi Appliance
Submit your organization details below. We review every order personally, confirm your staging requirements, and email you a direct payment link or corporate invoice. Appliances are benchmarked, staged, and dispatched upon receipt of payment.
Fulfillment process
From order request to running deployment
-
Order Review & Verification
We review your order details, confirm your delivery destination, and verify your network staging requirements.
-
Invoice & Payment
We email you an order confirmation with a secure Stripe payment link for corporate card payment, or a PDF invoice with ACH and wire instructions.
-
Burn-In Testing & Dispatch
Upon receipt of payment, your dedicated appliance undergoes a 24-hour hardware burn-in, fresh OS flashing, model verification, and ships via tracked courier.
-
15-Minute Plug & Play Go-Live
Plug the appliance into power and your isolated management network. Open the local console in your browser to begin running governed AI agents immediately.
Try the live demo first
Test the live demo and evaluate the runtime governance engine directly:
Review our open-source codebase on GitHub or read the Developer Guide. RunSAFi exists for teams who want turnkey physical hardware, pre-configured models, and production assurance.
Service Level & Assurance
What your $1,440/year assurance subscription covers
The hardware is yours permanently. The assurance subscription ensures your appliance remains hardened, compliant with changing AI regulations, and armed with the latest model weights.
Bi-Monthly Stable Releases & TCB Attestation
Systematic updates every two months with cryptographic TCB verification, proving your running governance core matches published SAFi releases for auditors.
Hardware-Optimized Model Upgrades
Benchmarked open-weight model upgrades delivered as newer, higher-quality weights emerge that fit the appliance's 64GB unified memory budget.
Direct Engineering Email Support
Email assistance directly from the engineers who built SAFi to resolve agent configuration, environment integration, and system-related issues.
Regulatory Rule Updates
Personalized assistance with policy updates as the regulatory environment changes.
Backups and Disaster Recovery
Automated local backup routines and tested restore procedures to safeguard your cryptographic audit ledgers and agent history.
OS & Kernel Security Hotfixes
Priority delivery of critical security patches and hotfixes to keep the underlying Linux operating system and the SAFi architecture secure.
Enterprise & Multi-Node
Beyond a single appliance
Need SAFi deployed across a multi-server datacenter cluster, private cloud VPC (AWS GovCloud / Azure), or high-availability Kubernetes setup? We provide scoped enterprise architecture services.
- Multi-node clustering. Distribute inference and governance across multiple GPU/APU nodes.
- Custom policy authoring. Direct assistance translating complex institutional bylaws into Synderesis charters.
- Bespoke MCP tool connectors. Connect agents securely to proprietary internal databases and ERPs.
- Custom cloud VPC. Deploy inside your private AWS GovCloud, Azure, or GCP tenant with automated Terraform scripts.
- Dedicated 24/7 SLA. Critical severity response times with emergency escalation channels.