RoadAI

Sovereign intelligence. Local-first. Your hardware. Zero cloud dependency.

52
TOPS
5
Pi Nodes
2
Hailo-8
27
Agents
$0
Per Month

Fleet Nodes

Alice
Gateway — Pi-hole, PostgreSQL, Qdrant, 51 services
Cecilia
AI Engine — Ollama, 4 models, MinIO, 457GB NVMe
Octavia
Architect — Gitea (239 repos), NATS, Docker
Lucidia
Web Fleet — 334 apps, nginx, PowerDNS
Aria
Mesh Node — WireGuard relay, compute

Models

Llama 3.1 8B

General purpose. Runs on Cecilia. 35 tok/s.

Mistral 7B

Fast inference. Code and reasoning. 40 tok/s.

Phi-3 Mini

Lightweight. Runs on any Pi. 50 tok/s.

DeepSeek Coder

Code generation and debugging. 30 tok/s.

Hailo Vision

Real-time object detection. 26 TOPS per chip.

Nomic Embed

Text embeddings for RAG and vector search.

Sovereignty

Local-First

All inference on your hardware. Nothing leaves your network.

$0/mo Inference

vs $200+/mo for cloud AI APIs. One-time hardware investment.

Audit-Logged

Every AI action recorded locally. Full transparency.

Any Device

Raspberry Pi, laptop, desktop. Sovereign AI everywhere.