ON-PREMISES DEPLOYMENT
Bring voice AI into your own infrastructure.
Deploy agent workflows within your organization’s environment, with control over data placement, enterprise-system access and operations. Choose local services or approved external AI providers as part of the configuration.
Operate within your organization’s environment.
On-premises deployment gives organizations a way to run core platform services alongside their existing systems. Define where agent execution, workflow state, knowledge and connector access belong, and who manages the environment.
Plan access controls, customer keys, audit records, monitoring, updates and recovery as part of the deployment. Hosted speech or reasoning providers can remain external; an on-premises runtime does not automatically make the entire service isolated.
SECURITY THROUGH THE STACK
Identity at entry
OAuth2 / OIDC · RBAC · tenancy
Keys under your control
Customer keys · Vault · scoped credentials
Rules around each action
Consent · PII guardrails · permitted tools
A record of the operation
Audit · agent versions · human decisions
KubernetesRuntime placement · not a live server monitorOrchestration · state · campaign engine
Permitted enterprise actions
PostgreSQL
Redis
OpenSearch
Terraform
GitLab
OpenTelemetry
GrafanaNetwork routes, identity, data placement and provider connections must be agreed for the chosen cloud configuration. Controls shown here are from the reference architecture; implementation and security evidence are reviewed for your deployment.
What we need to plan together.
START WITH ONE WORKFLOW
Bring your SOP. Let’s map the first process.
Define the task, connect the systems and agree what a useful pilot should prove.