TruSpark™ runs inside your environment, not ours. There is no hosted component and nothing to send us. What you build stays with you in open, standard formats, so your knowledge remains usable instead of merely retrievable. Most of what follows is a property of how the platform is built, not a policy you have to trust us to honor.
TruSpark ships as self-hosted software — SDKs, a native library, and an on-premises engine that run inside your own infrastructure. There is no cloud service and no hosted component: your ontologies, your data, and the software you generate never leave your boundary. Data residency is not a policy you have to trust us to honor. It is a property of how the platform is built.
Nothing routes through a Reasonics service. The platform is yours to run.
We never store or process your data. It stays inside your infrastructure.
Because it runs where you deploy it, where your data lives is your decision, never ours.
Your data, your ontologies, and the SDKs and artifacts TruSpark generates from them belong to you, and we do not hold your knowledge hostage. Everything the platform produces is expressed in portable, open formats, so what you create is yours to keep, move, and use, with or without Reasonics.
Your ontologies, your generated code, and your proof certificates all work on their own, no proprietary format to decode, and no dependency on us to use them. The one exception is the graph schema our runtime compiles from: it's readable and open, but turning it back into running code takes our compiler. If you ever leave, you leave with everything you brought and everything you built.
Access is governed by multi-factor authentication and both role-based and attribute-based access control. Licensing is enforced with cryptographically signed, machine-bound tokens, and the native interop layer uses authenticated encryption over a typed binary channel. We follow certification-ready practices throughout.
Multi-factor authentication with role-based (RBAC) and attribute-based (ABAC) authorization.
Ed25519-signed tokens bound to the machine they run on.
AES-256-GCM over a typed binary channel, not ad-hoc JSON.
Engineered so that when your requirements call for certification, we are prepared to support it.
TruSpark includes operational telemetry that we can tailor to each deployment, and we provide reporting on exactly what it captures, so what is collected is always transparent to you.
Security and privacy keep your data safe. Correctness is a separate guarantee, and it is one you can settle without us. A proof certificate is a portable record of the reasoning behind a result, written in published formats that independent tools can read. Your auditor can re-run the check on their own machine, years later, with Reasonics gone from the picture entirely, and reach the same verdict or find that it fails.
This is what separates evidence from assurance. A vendor claiming its system is correct is asking for trust. A certificate anyone can re-check is not a claim at all. It either verifies or it does not, and we have no say in which.
The chain of trust behind those certificates, the standards, the theorem provers, and the formats the evidence is written in, is set out in full on The Architecture of Trust.