Can I run iDBQuery with my own on-premise AI model?

For organisations that cannot send anything to an outside service, iDBQuery supports running the language model on your own hardware.

• Local inference - the model runs inside your network, so prompts, schema details, and results never leave your boundary.
• Air-gap ready - a locally hosted model is what makes a fully offline deployment possible, with no callbacks or telemetry.
• Same experience - you still ask questions in plain language, get the SQL written for you, and receive answers cited back to source rows; the difference is purely where the model runs.
• You control the environment - the model lives in your patching, monitoring, and hardware lifecycle alongside the rest of iDBQuery.

For example, a hospital or defence group can run iDBQuery with an on-prem model so that even the text of a question about patient or mission data stays internal. This is a deliberate architectural choice rather than a workaround: iDBQuery is designed so the intelligence layer can sit inside your walls, which is stronger than any policy promise about a remote service. It pairs with query-in-place data access, RBAC, and encryption in transit and at rest to keep the entire pipeline local. We help size the compute for the model during planning and validate that no external calls are made. Contact us to discuss a bring-your-own-model deployment.