FAQ
Frequently asked questions
Straight answers for on-prem AI, deployment, costs, and governance.
How this differs from online AI
Why is there no cloud trial? +
ModelLinq is built for on-prem deployment, which requires access, data, and governance setup. We provide online meeting demos or PoC.
Why not emphasize model size? +
Model size is not the only performance factor. Governance-driven orchestration and task success within security and cost constraints matter more.
Is on-prem AI weaker? +
Our focus is compliance and fit for target industries, with governed workflows, data protection, and controlled outcomes.
Can we use our own models? +
Support is limited and depends on model size and available server compute.
Cost and deployment concerns
Do we have to buy hardware? +
You can use existing servers or procure new ones. We help assess specifications with integration partners.
How is cost calculated? +
Costs are based on project delivery and maintenance contracts. Pricing is fixed and predictable, not usage-based.
Is it suitable for SMBs? +
If you need high security or predictable costs, smaller teams can start with a right-sized deployment.
How long does deployment take? +
It depends on scope and environment. We often start with a small PoC and then expand gradually.
Security and governance
Will data leak out? +
Data stays on-premises, with access control and masking to reduce leakage risk.
Is there auditing? +
Every model change, inference, and usage action is fully auditable.
Can we restrict model sources and languages? +
Yes. You can whitelist model sources, versions, and allowed languages to match internal policies.
Can it run in air-gapped networks? +
Yes. ModelLinq supports deployment in isolated environments with no external connectivity.
Still have questions?
Tell us about your environment and we will map the right on-prem plan for your team.