A pharmaceutical facility shouldn’t hand its precise equipment data to a random cloud app — and with SEU, it doesn’t have to. The architecture is local-first and single-tenant by design, so trust isn’t something you extend on faith; it’s how the product is built.
Your plant model — every equipment parameter, every connection — is stored locally on your device. SEU does not operate a server that holds your facility data. Close the tab and the model is on your machine, not ours.
For a deployment, SEU runs inside your own environment — your cloud or on-premises. Even with accounts and storage enabled, your data sits behind your walls and your IT controls, never a shared multi-tenant system.
The only time anything leaves your machine is if you choose to use the AI assistant — and then only the specific request you send, to Anthropic’s API, which does not train on it. Every calculation runs locally; nothing is uploaded in the background.
There is no analytics harvesting of your model, no resale, no third-party brokering. The audit you build is yours. We make money from the software, not from your data.
Equipment, parameters and connections — entered and computed in your browser.
The model and results persist in your device’s local storage. No server of ours receives it.
You upload a meter file (processed locally) or use the opt-in AI assistant — only the request you send.
The browser is told to load scripts, styles, fonts and data from this origin only. Injected content has nowhere to call out to.
No analytics, tag managers, external fonts or trackers. Every byte is served from one origin — which is what makes a policy that strict possible.
HTTPS everywhere, with HSTS preload — browsers refuse to fall back to plain HTTP.
frame-ancestors none and X-Frame-Options DENY block click-jacking; the app is never framed inside another site.
The optional assistant is rate-limited and size-capped, times out in 30 s, validates every input, and is never cached. The model key lives only on the server.
A standards-based security.txt (RFC 9116) and a published disclosure policy — a researcher always has a clear, private way to report.
These are response headers and code you can inspect — view source, read the security.txt, or check the policy on any page. SEU runs no analytics or telemetry today; if usage analytics are ever added they will be opt-in and counts-only — never plant data.
Audit-trail and data-integrity mindset throughout the model and report.
Built for the constraints of classified, sterile manufacturing.
EU data residency available; personal data minimised and consented.
The energy-review structure the whole engine is organised around.
SEU is engineered to support these frameworks; formal certification is part of an enterprise deployment, scoped with your quality and IT teams.
A deliberate line between what is true in the open demo now and what an enterprise deployment adds — we don’t claim a certification that isn’t in place.
We’ll walk your IT and quality teams through the architecture and a single-tenant deployment.