LLM inference engine
3B–8B quantized models via Ollama / llama.cpp. Optimized per hardware target with latency <200ms first token.
Sovereign AI stack for edge, on-prem, and OEM
A private system that answers domain questions, reasons within bounds, and safely executes validated actions — offline and in real time.
Capability grid
Production-oriented layers — not a single-model chat wrapper.
RAG layer
Private document ingestion, local vector DB (FAISS/Qdrant), top-k retrieval with reranking for grounded answers.
Command layer
Structured JSON intents mapped to validated actions — the differentiator that makes Sovra deployable in production.
Application layer
Prompt templates, orchestration, guardrails, and deterministic fallbacks for reliability.
API & UI
REST API for OEM integration plus web chat UI. Phase 2 FastAPI backend; Phase 1 demo on shared hosting.
Monitoring
Local latency tracking, error logging, and privacy-preserving usage analytics — no cloud telemetry required.
Deployment modes
Edge
Raspberry Pi & Jetson
Offline-capable assistants, IoT intelligence, automotive prototyping.
On-prem
Mini PC & GPU server
Team knowledge bases, meeting intelligence, enterprise RAG at scale.
Automotive
OEM integration path
In-vehicle voice, owner manual Q&A, driver support — custom ECU modules.
Choose your hardware
Map platform capabilities to a certified stack with live pricing.