Privybox, integrated with your systems

Every Document. Every System. 100% Local.

Software, systems-integration consulting, and deployment, combined — tailored to your organization and deployed on infrastructure you control.

💬

Chat With Your Documents and Systems

Query your internal documents and live business systems in natural language — all processed locally, with nothing sent to the cloud.

🔒

Fully Local, by Design

Embeddings, retrieval, and inference all run inside your own infrastructure. Your data never leaves your network.

🏢

Department & Role-Based Access

Control access by department, with two permission levels — staff and manager — so people only see what they're meant to see.

🔌

API-First, With a Web UI

Every capability is exposed as an API for your own integrations, plus a ready-to-use web interface out of the box.

Our Implementation Methodology

Who it's for

Privybox Enterprise is built for small and mid-sized businesses whose core information needs to stay inside their own infrastructure — not sent to a public cloud AI service.

How it works

No enterprise AI deployment can be fully "out of the box" and still fit every business. So our methodology rests on three things:

Pre-built core

The application itself is already built, tested, and ready to deploy — not built from scratch for each customer.

Industry-standard configuration

Customer-specific setup is done against standard interface patterns — like OData $filter queries or REST/JSON schemas — not bespoke reverse-engineering.

Minimal custom development

True custom code is kept to the small number of cases that standard configuration genuinely can't cover.

Timeline

On-site implementation: five business days, maximum

One business week from kickoff to go-live — fully installed, tested, and ready to use by the time we leave. Everything that can be prepared remotely is finished before we arrive.

1 A person filling out the online intake form at a computer

Web Intake

Basic company info

2 A team discussing scope and pricing around a table

Scope & Contract

Pricing + deposit

3 A team at workstations in front of a large screen showing a remote site feed

Remote Customization

Interfaces + configuration

4 Hard-hatted engineers reviewing a blueprint on site

On-site Week

5 business days, start to finish

Implementation Testing Go-Live + Training
5 The team celebrating a completed project

Final Payment

Project closed

Five stages from first contact to go-live. Scope, pricing, and the contract are all settled before any on-site work — and once we're on site, installation through knowledge transfer is capped at five business days.

After go-live, support continues remotely while your data stays on your infrastructure — we can help diagnose issues and ship updates without ever touching your underlying data.

Technical Architecture

Four layers, from the apps your team uses down to the systems you already run — all on your own infrastructure.

Privybox Enterprise technical architecture: four layers from the user-facing applications down to your data sources Layer 1 · ApplicationLayer 2 · OrchestrationLayer 3 · Data AccessLayer 4 · Data Sources Userdirect query — bypasses search & toolsLayer 1 · ApplicationChat Assistant UI — natural-language Q&AStandard Reports UI — pre-built analyticsLayer 2 · OrchestrationAPI Layer — request orchestrationLLM — on-prem inference, invoked by the API layerLayer 3 · Data AccessVector Database — semantic search (RAG)Reports Engine — fixed queries, direct to sourceMCP Interface — tool callsLayer 4 · Data SourcesDirectory Files, Master Data, Transaction & Auxiliary System interfaces
API requests reach your data either through the vector database (semantic search) or an MCP tool call — except the Reports Engine, which queries Master Data and Transaction interfaces directly, bypassing both.