The terms we use
Short definitions, written once and linked from everywhere else.
Denizen Blu One
Denizen Blu One is the customer-facing product: one subscription that spans chat and inference, applications, notebooks, training and managed compute, on the web and in native apps. You declare what to run and Denizen Blu chooses the execution path.
Conduit
Conduit places each request on suitable infrastructure so providers, regions and silicon stay out of the buying decision. You never choose a GPU or a region to use a model.
Maestro
Maestro is part of the layer that keeps providers, regions and silicon out of the buying decision, alongside Conduit and Orchestra.
Orchestra
Orchestra is part of the layer that keeps providers, regions and silicon out of the buying decision, alongside Conduit and Maestro.
KaaS (Kubernetes as a Service)
KaaS gives organizations managed Kubernetes environments, so teams run containerized workloads without sourcing or operating the underlying infrastructure.
SLaaS (Slurm as a Service)
SLaaS provides managed Slurm scheduling for model training, high-performance computing and large distributed jobs, with the cluster operated for you.
Seat
Instead of asking you to shop for GPUs, Denizen Blu sizes a workload into seats. Plans are flat-priced, and Enterprise is priced per seat.
Fair use
Tokens are used as telemetry and fair-use capacity, not as billing. The fair-use policy explains the limits that keep flat plans fair for everyone.