Privalyx
  1. How it works
  2. High-Velocity Compute
Local-first AI

Software Architecture

Private local computing for project controls, AI and secure collaboration. Optional cloud services connect only where the workflow requires them.

Deployment overview

Where project data is processed

User-controlled environment

User PC

On-premises / local environment

Private boundary
Privalyx App

Application layer

User InterfaceProject workspace and controls
Communication HubTeam and service connections

Private local computing

AlgorithmsProject controls functions
RAG CorpusProject and domain knowledge
Local-first AIRouting and guardrails

Secure local storage

Project data and settings
Local processing output
AI output
Encrypted syncMinimal report outputOptional connections

Privalyx Platform

Secure cloud services on Microsoft Azure

Zero-knowledge Data SyncEncrypted sync, zero dependency
SSO and LicensingIdentity and entitlement services

Microsoft Foundry

Transcient analysis

Zero project knowledgeProject schedules stay local
Zero data storageNo chat history or report output retained
Guarded AI outputControls and guardrails remain local

Project data, local processing output and AI output stay in the local environment. Optional services receive only the data required for their specific function.

At a glance

Key points

  • 01

    Compute locally

    Project controls, algorithms and AI routing run in the Privalyx app in your environment.

  • 02

    Keep project data local

    Project data, local processing output and AI output stay in storage under your control.

  • 03

    Connect services as needed

    Sync, SSO and licensing are optional services; sync uses zero-knowledge encryption.

  • 04

    Limit cloud AI processing

    Only necessary output is sent for optional Microsoft LLM (Foundry) processing; chat history and report output are not retained.

Architecture layers

What runs where

01

User application

The interface and communication hub run in the Privalyx application on the user’s PC or local environment.

02

Private local computing

Project-control algorithms, the RAG corpus, and local-first AI routing and guardrails work with project context locally.

03

Secure local storage

Project data and settings, local processing output, and AI output are held in local storage under the user’s control.

04

Microsoft Foundry services

Privalyx Platform provides zero-knowledge data sync, SSO, and licensing. Microsoft Foundry processing (when inquired) receives only necessary output, with zero data storage.