K-AI setup
K-AI is configured separately from the GenAI Gateway, which connects workflow nodes to external providers. By default, K-AI routes requests through Knime's managed AI infrastructure. Organisations that need data residency control or want to use their own model configure K-AI's backend directly — see Bring your own model.
What you can configure
AI provider — choose between Knime-managed models or connect your own OpenAI or Azure OpenAI backend. See Bring your own model for the supported providers and setup steps.
Data residency — when you use your own provider, prompts and responses travel through your own cloud infrastructure and never leave your chosen region. With Knime-managed models, data is processed according to Knime's data processing agreement.
Provider compliance — in Knime-managed AI mode, OpenAI processes requests as Knime's sub-processor. You and your authorized users must comply with OpenAI's usage policies, in addition to Knime's own terms. When you connect your own provider, you contract with that provider directly and comply with its terms instead.
Usage policies — admins can restrict which K-AI features are available to users in the account: for example, disabling code generation or limiting which models are accessible.
Audit logging — all K-AI interactions are logged. Admins can review usage through the audit log — see Audit trails.
Where to configure
- K-AI backend and provider settings — see Bring your own model.
- GenAI Gateway (for workflow nodes, not K-AI) — Admin panel → GenAI Gateway. See GenAI Gateway.
- Bring your own model → — data handling, model selection, and compliance considerations
- What K-AI can do → — capabilities overview for workflow builders