Our vision for AI partnerships
The current shape of AI is mostly extractive: your data trains someone else’s model, your attention feeds someone else’s product, and the intelligence you talk to lives on a server you’ll never see and don’t control. That’s not a criticism of any one company — it’s just the default architecture the whole industry grew up inside.
We think a different relationship between people, their AI, and the model labs that build the underlying intelligence is possible — one where the tool is yours, the data stays yours, and the model providers are partners you choose to bring in, not gatekeepers you have no choice but to go through. That’s the actual reason this page exists: not to court a deal, but to say plainly what we think this should look like, and why it matters that someone builds it this way.
This isn’t built against the model labs. It’s a free, local utility that can, at the user’s choice, route work to them — on the user’s own subscription. This is an invitation, not a demand.
The tool is given away
The Foundation’s core tool is free. It is not sold, not subscription-funded, and not paywalled. It is a public gift, not a business that charges for access.
Local-first by design
It runs on your own device, works offline, keeps your data on your machine, and avoids cloud dependency and data extraction. That is the spine, and it does not change.
Your cloud, your choice
If you already pay for a cloud model — Claude, Gemini, ChatGPT — you may optionally connect your own subscription as an added “cloud brain.” You authenticate directly with that provider; the Foundation never owns, resells, or marks up their service. Used this way, a model lab becomes an optional upstream provider to a free public utility — not the owner of it.
Aligned incentives
This makes the relationship additive, not adversarial. Providers could gain direct demand and subscription revenue from people who value this tool; the Foundation keeps its sovereignty, governance, and mission. The interests can point the same way — which is rare, and worth building toward deliberately rather than assuming it happens on its own.
Why this matters
If the only way people get access to powerful AI is through accounts they don’t own, running on infrastructure they can’t see, trained on data they didn’t consent to give — that shapes what kind of relationship humanity ends up having with this technology, quietly, by default, without anyone choosing it on purpose. We think the alternative is worth building even before anyone asks us to: something local, something free, something that treats a model provider as a partner a person opts into, not a toll booth they can’t avoid.
An open invitation
Conversation with model providers and institutional allies about how this could grow — co-marketing, capacity, shared standards — is welcome. Here is the architecture; here is why it could be upside for everyone; here is what partnership could look like. Reach out via the contact page.
Sovereign Grace Foundation 