Private AI help for teams that care where data goes

Use AI. Keep your data in your hands.

Learn how to test open-source and open-weight models. Build an agent on your own machine or server. See what it can read, what it can save, and what it can send.

LOCAL MODELS · OPEN SOURCE · OPEN WEIGHTS · PRIVATE SERVERS · DATA RULES · AGENT TOOLS · AGENT MEMORY

Cloud AI is easy to start

The hard part is knowing where your data goes.

Your team may paste client notes, plans, code, or files into AI tools. You need clear facts before you can trust that work.

01

Private work may leave your hands.

You may not know where a tool sends a file, how long it keeps it, or who can see it.

02

A vendor can change the deal.

Price, limits, model access, and rules can change. Your key work should have a safe plan B.

03

An agent can reach too far.

An agent may read files, call tools, or save facts. Each power needs a firm rule.

04

“Open” can mean many things.

Open source and open weights are not the same. The license and files tell you what you can use, study, change, and share.

The Open Source Initiative sets a high bar for the term open-source AI. Many models share their weights but not all of the parts needed to study and remake the model. We name that gap. Read the definition ↗

Two ways to learn

Start with one private AI use case.

We map the data, pick a model, set rules, and test a small system. You learn each choice as we work.

Private · built for your case

Sovereign AI working class

$297 / 60 minutes

Bring one task that uses private data. We will map a safe first test. We may run a model, sketch the full system, or set rules for an agent.

  • List the data the task needs
  • Pick an open-source or open-weight model
  • Choose your machine, server, or private host
  • Set rules for tools, logs, and memory
Plan my private AI system →

Tell us the data you need to protect and the job you want AI to help with.

Weekly · set lab

Sovereign AI group lab

$29 / live class

Each lab builds one small part of a private AI stack. You can follow the steps and ask how they fit your own case.

  • See the model and task before you book
  • Run each step with the group
  • Learn the cost and speed trade-offs
  • Keep the setup notes and safety checks
See the next private AI lab →

Pay for the labs you want. You do not have to join a plan.

Four lab topics

Build control one clear step at a time.

We post the date, model, tools, and goal before each lab.

LAB 01 · MAP

Draw where your data goes

List each file, tool, model, log, and person in the path.

LAB 02 · RUN

Run a model on your own machine

Use a tool such as Ollama. Test speed, fit, and limits.

LAB 03 · SEARCH

Ask questions about private files

Build a small search tool. Keep the files in a place you choose.

LAB 04 · ACT

Give an agent one safe job

Set a short tool list. Save only the memory it needs. Keep a stop switch.

What you take home

Know where your data goes and who controls each part.

Know what can stay on your machine.

Know what each model license lets you do.

Set tight rules for agent tools and memory.

Keep a clear path away from one vendor.

Quick answers

Before you book.

What does sovereign AI mean here?

It means you make the key choices. You choose where the model runs, where data is kept, who can use it, and which tools an agent can call.

Are open source and open weights the same?

No. Open weights give you the trained model weights. True open-source AI also gives broad rights and the key parts needed to study, change, and share the full system. We check the license and files for each model.

Do I need a large computer?

It depends on the model and job. Many small models run on a laptop. Large models need more power. We can map the need before you buy gear.

Will this make my system private and compliant?

No class can prove that on its own. Privacy depends on the full setup and how your team uses it. Laws and rules also vary. We help you map the system and set clear controls. This is not legal or audit advice.

Do I need tech skills?

No. We can start with a plain data map and a small local model. If you have a tech team, we can go deeper.

Own the key choices

Build AI you can see, test, and control.

Start with one private task. Learn how to pick the model, place the data, and set safe rules for the agent.