A model endpoint is any inference API or serving endpoint that NeuralVault connects to for monitoring or testing. Each distinct URL or deployment slot counts as one endpoint, regardless of the number of requests processed.
Yes. You can upgrade at any time and the change takes effect immediately. Downgrades take effect at the next billing cycle. There are no penalties for plan changes.
We offer a guided 14-day proof-of-concept for teams on the Professional plan or above. Contact our sales team to get started. We don't offer self-serve trials because meaningful AI security testing requires a proper onboarding session.
Enterprise plans are quoted annually based on number of model endpoints, probe scenario volume, and compliance reporting requirements. Annual and multi-year commitments include significant discounts. Contact our sales team for a custom quote.
Starter and Professional customers are hosted in the US region by default. Enterprise customers can choose from US, EU, or APAC data residency to meet regulatory requirements. NeuralVault does not retain your model weights or inference payloads beyond the configured log retention window.
NeuralVault generates audit evidence packages mapped to SOC 2 Type II, ISO 42001, EU AI Act (Title III), NIST AI Risk Management Framework, and GDPR Article 22. Additional frameworks are available on Enterprise plans.
Not sure which plan is right for you?
Talk to our team. We’ll help you find the right fit for your model portfolio and compliance requirements.
Neural