What private AI actually means.
Private AI means the model, and the system that feeds it your data, run on infrastructure you control instead of a third-party AI service. Your team asks questions in plain English and gets answers drawn from your own documents and database, not from the public internet.
For a lot of businesses that's the difference between "we can't put that in ChatGPT" and "we use it every day." Contracts, client records, internal policies and support history stay where they are.
How it works: retrieval-augmented generation.
The technique behind it is called retrieval-augmented generation, or RAG. When someone asks a question, the system first searches your documents and database for the relevant passages, then gives only those passages to the model to write the answer.
No retraining
The model doesn't have to be trained on your data. It reads what's relevant at the moment it's asked.
Always current
Add or change a document and the next answer reflects it.
Grounded answers
Responses come from what your business actually knows, not the model's general guesswork.
What teams use it for.
- Triage and categorize email, so the messages that matter reach the right person first.
- Manage calendars: find times, schedule meetings and keep everyone's day in order.
- Search your documents, from policies and manuals to contracts and procedures, and get answers in plain English.
- Ask your database questions, like "which accounts haven't ordered in 90 days?"
- Sort and categorize records that your team currently tags by hand.
Built on Hermes Agent, open source and self-hosted.
We build private AI as custom agents on Hermes Agent, the open-source (MIT) agent framework from Nous Research. It runs on your own servers, keeps a persistent memory of your business, and works with open-source models you host yourself, so sensitive data never has to leave. When a task needs a commercial model, you decide what goes out and what doesn't.
Why this needs engineers, not a subscription.
The model is the easy part. The work is connecting it to the systems you already run, respecting who's allowed to see what, and keeping it working after launch. That's what we've done since 2008: integrations, pipelines, monitoring, and still being there when something breaks. If owning your data matters to you beyond AI, see our self-hosting and data ownership services.
Private AI questions.
Does my data leave my infrastructure?
Not unless you want it to. Hermes Agent can run entirely on infrastructure you control, with open-source models you host. If you choose a commercial model for a task, you decide what data it sees.
What's the difference between private AI and uploading files to ChatGPT?
Where the data goes and who controls it. A private deployment keeps your documents and the model on your infrastructure, and connects to your live systems instead of relying on one-off uploads.
What do I need to get started?
A set of documents or a database your team wants answers from, and a team that will use it. A good first step is our 2-week AI Sprint (from $6,500): we map where AI pays off and build one working prototype on your own data.
Is private AI available now?
It's in pilot with a limited number of businesses. Get in touch to ask about a spot.
Limited pilot spots. 20 minutes with an engineer, not a salesperson.
Ask about the private AI pilot.
Tell us what your team keeps searching for. We'll tell you honestly whether private AI can answer it. Email hello@netlandish.com, call +1 562-318-0100, or use the button below.