Services Ver en español

Private AI Server

Every question you type into a hosted AI service leaves your building, and so does every document you attach to it. For a firm handling contracts, medical records or client files, that is the whole objection, and it is a reasonable one.

This is a model running on a machine in your office. It answers questions, summarises documents and connects to your existing tools through an API. Nothing it reads goes anywhere, because there is nowhere for it to go.

What it costs

How it works

  1. We size it to the workA small model answers questions and holds a conversation. A medium one reads and summarises documents. A large one helps with research, writing and code. The tier is the model, and the model is the cost.
  2. We install and connect itOn site, on hardware that stays with you, with an API so your existing software can call it rather than somebody retyping into a chat window.
  3. It keeps runningOngoing support is a monthly add-on covering updates, maintenance and performance work. Without it the machine still runs; it just stops improving.

Who this is for

Who this is not for

You want the best answers available
You will not get them here. A model that runs on one machine in an office is smaller than one running in a datacenter, and it answers accordingly. You are buying privacy, and it costs capability.
You have no policy reason to keep the data in
Then this is an expensive way to buy something available by subscription, and we will say so before you spend $800 on it.
You want it to run the business
It answers and summarises. Quotes, invoices, appointments and service reports are a different system, and pretending one machine does both ends with neither done well.

Worth having alongside