UnitVane documentation
Set up UnitVane, KubeOn, UnitVane AI and the GPU optimizer. Choose a product, connect its data and understand the controls before rollout.
Choose your solution
UnitVane
Bring every cloud bill into FOCUS, map it to your business, and build your own dashboards.
Start with UnitVaneKubernetes
Connect a cluster and its bill. Allocate cost, review savings and report to each team.
Kubernetes quickstartAI tokenomics + gateway
Connect provider costs, understand tokens and capacity, then control model requests.
Start with UnitVane AIGPU optimization
Measure every GPU-hour, then let the opt-in optimizer share, scale, burst and switch off GPUs.
Start with KubeOn GPUUnitVane is a SaaS platform with three products. UnitVane covers every cloud bill and needs read access to billing only. KubeOn adds a read-only agent per cluster, pod-level allocation and change proposals, optional write access for automated savings, and the GPU optimizer. UnitVane AI covers model spend, with an optional managed gateway.
Start here
- Quickstart: from zero to the first allocated costs.
- Architecture: what runs where, and how data moves.
- Install the agent: Helm values for every platform.
- Connect AWS billing, Azure, Google Cloud or on-premises rate cards.
UnitVane
- Get started with UnitVane: connect a source and open your first dashboard.
- Connect data sources: AWS, Azure, Google Cloud, OpenCost and FOCUS files.
- Business mappings, dashboards and your BI tool, budgets, anomalies and savings.
UnitVane AI
- Gateway quickstart: one endpoint for every model.
- Connect AI providers: Azure OpenAI, Bedrock, Vertex AI and APIs.
- Provisioned capacity sizing and tokenomics methodology.
KubeOn GPU
- KubeOn GPU overview: what each capability needs and the order to roll it out in.
- Connect GPU metrics, install the GPU optimizer and the GPU policy reference.
- On-premises rate cards: GPU-hour rates for servers you own.