Metered Usage and Billing
Ardor’s pricing is based on a metered usage model, ensuring that you only pay for the actual runtime of your AI Agents. When an AI Agent is not in use, no charges apply, making the platform cost-efficient and scalable.
How Metered Usage Works
Ardor tracks AI Agent resource consumption in real-time and charges only for the duration in which an agent is actively running. This applies to:
- vCPU Usage
- RAM Usage
- vRAM Usage
- Storage Usage
If an agentic solution is idle or turned off, it does not consume CPU, RAM, or vRAM, and therefore incurs no charges for compute resources. However, persistent storage costs still apply as long as data remains allocated.
Billing Breakdown
Compute Resources (CPU, RAM, vRAM) – Pay Only for Active Use
| Resource | Billing Method | Charge Applies When… |
|---|---|---|
| vCPU | Per second per vCPU | The AI Agent is actively running tasks or processing data. |
| RAM | Per second per GB | Memory is allocated while the agentic solution is in use. |
| vRAM | Per second per GB | Video (GPU) memory is used for processing AI workloads. |
Once an AI Agent stops running, metered billing for compute resources pauses automatically, and no further charges apply.
Storage – Persistent and Billed Monthly
| Resource | Billing Method | Charge Applies When… |
|---|---|---|
| Storage | Per GB per month | Data, models, or logs remain stored on the platform. |
Storage is billed regardless of whether an AI Agent is active or inactive, as long as the data is allocated.
Ardor’s metered billing model ensures that businesses only pay for what they actually use, making cloud agentic solutions scalable and cost-effective.