Skip to content
Let’s talk

Open models on your own hardware or private cloud, answering questions over your documents for one workflow.

Self‑hosted AI

Timeline, price and terms

Timeline

4 to 6 weeks, plus hardware delivery if you buy hardware

Price

Pilot
$12,000

We recommend it only when the workflow diagnostic shows it pays.

Payment

50% before we start, 50% on delivery.

Ownership

The code, the prompts and your data are yours. NDA before the first call.

What you get

  1. Open models running on your hardware or private cloud.
  2. Search and answers over your own documents (RAG).
  3. One workflow, running from start to finish.
  4. Your documents stay on your own systems.

How it runs

  1. Diagnose

    Before the pilot

    The workflow diagnostic tells us whether self-hosting pays for your case.

  2. Set up

    Weeks 1 to 2

    We install an open model on your hardware or private cloud.

  3. Connect

    Weeks 2 to 4

    We index your documents so the model can answer from them.

  4. Run

    Weeks 4 to 6

    Your team uses it for one workflow.

Questions

How long does the pilot take?

4 to 6 weeks once your hardware or private cloud is ready. If you need to buy hardware, add the supplier’s delivery time. We tell you what to buy before you order.

Is the hardware included?

No. Hardware and cloud hosting are billed to your own accounts at cost. We add no markup.

Why only after a workflow diagnostic?

Running your own models adds hardware and upkeep. We recommend it only when the diagnostic shows it pays.

What is RAG?

Retrieval-augmented generation. The model looks up the relevant passages in your documents before it answers.

Which models do you use?

Open models. We pick one that fits your hardware and the workflow.

Reading on this

Next step

Start withtwo minutes