Privacy Policy

Last updated July 9, 2026

This policy reflects how Garnet works today; it may be updated. Questions? Email privacy@garnetgrid.com. The short version: Garnet builds a private AI that runs on hardware inside your building. The system itself has no telemetry and does not phone home. Pull the network cable and it still answers. Once a build is handed over, we hold no standing access to it. This website collects only the contact and scoping details you choose to send us when you ask us to scope a build. We don't sell your data, and your work never trains anyone's model.

A Garnet build is a private AI that runs on hardware in your building. Your prompts, your documents, and the AI's replies are processed on that hardware and stay there. They are never sent to us or to anyone else. The system carries no telemetry, no analytics, and no phone-home channel; it works with the internet unplugged. Once the build is delivered and handed over, we retain no standing access, no remote login, and no back door into it. What the system does after handover is yours alone, and we could not see it even if asked.

Garnet is not a downloadable app, and there is no on-page chat that quietly routes your words to a cloud AI. Two things do happen here: you can tell us about the confidential work you need the system to handle and we scope a build; and you can buy a fixed-scope engagement or a document product through a Stripe checkout link. Stripe processes the payment and sends us your name, email address and what you bought; we keep that to deliver the purchase and for our accounts (see section 5). The rest of this notice is about those two interactions, not about a product you install yourself.

• Scoping-intake details. When you scope a build, we collect what you choose to share: your name, your organization, the general nature of the confidential work you want the system to handle, and how to reach you. This goes directly to us so we can prepare for and hold the scoping conversation. Share only what you're comfortable putting in an intake form. The deep detail belongs in the conversation, not the web form.
• Correspondence. If you email us or reply to a scheduling message, we keep that thread so we can follow up on your build.
• Basic website analytics. We use only minimal first-party analytics to understand aggregate traffic on this website — which pages are visited, and which links are followed. Those events are tagged with a random identifier kept in your browser's session storage, so that one visit reads as one visit rather than a scatter of unrelated hits. It is generated in your browser, contains nothing about you, is shared with no one, and is discarded when you close the tab — we cannot recognise you on a later visit. No third-party trackers, no advertising profiles. This is the website only; the systems we build for clients carry no telemetry at all.
• What your browser remembers. If you dismiss the newsletter box or sign up through it, we record that in your browser's local storage so it stops reappearing; and if a form submission fails to reach us, we keep it there until it can be retried. Those entries stay on your device, hold no identifier for you, and are never assembled into a profile. Clearing your site data removes them.
• Technical logs. Standard server logs, such as a pseudonymized IP hash for rate-limiting and abuse prevention.

To reach you about the build you asked us to scope, prepare for and hold the scoping conversation, coordinate delivery of a build you've commissioned, keep this website running and secure, and answer your questions. That's it. We do not use anything you send us to train any AI model.

We don't sell your personal information. We share it only with the service providers that operate this website and let us reach you: Cloudflare (hosting), our scheduling and email providers (to receive your intake and correspond with you). Each processes data only on our behalf. We may disclose information if required by law. The private AI systems we build for clients are not connected to any of these providers.

Stripe, Inc. — payment processing for purchases made through the checkout links on this site. Stripe’s own privacy notice governs the card details it collects; we never receive them.

This site sets no cookies. It uses two kinds of browser storage instead, both first-party, and neither one used to follow you between visits. Session storage holds the random per-visit identifier described in section 3, and is discarded when you close the tab. Local storage remembers that you dismissed or completed the newsletter signup, and holds a form submission that failed to send so it can be retried. Nothing stored here builds an advertising profile, goes to a third party, or keeps an identity for you from one visit to the next. If your browser blocks cookies or storage, the site still works.

Depending on where you live (e.g., the EU/UK under GDPR, or California under CCPA), you may have the right to access, correct, delete, or export the personal information we hold about you, and to object to certain processing. To exercise any of these, email privacy@garnetgrid.com and we'll respond.

We keep scoping-intake and contact information for up to 24 months after our last interaction (or sooner on request), then delete it. We use reasonable technical and organizational measures to protect what we hold. No method of transmission or storage is 100% secure. But the most sensitive thing, the work your Garnet system handles, never touches this website or our servers in the first place: it lives on hardware in your building, and after handover we have no access to it.

The site and service are not directed to children under 13 (or the equivalent age in your region), and we don't knowingly collect their information.

We'll update this page when our practices change and revise the date above. The service is operated by Garnet Grid Consulting LLC. Questions or requests: privacy@garnetgrid.com.