Using the Ganivra website

Illustrative product information

Dashboard values, customer names, cost examples, projected savings, and product-direction statements on this website are illustrative. They demonstrate the intended product experience and are not customer results or financial advice.

MVP accounts

You are responsible for your account credentials and SDK keys. Access, limits, pricing, and support may change while the MVP is being developed.

Acceptable use

You may use Ganivra to evaluate your own AI workloads. Do not disrupt the service, access another workspace, upload prohibited data, or represent illustrative material as verified customer data.

Cost estimates and GPU metering

Ganivra prices API calls from published or contract rates. Calls to models on your own GPUs have no per-call price, so Ganivra estimates one by splitting each UTC day's GPU cost across that day's calls. These figures are estimates for understanding your costs, not an invoice. By using GPU metering you accept these limits:

• Calls without a vLLM trace are split by tokens, which is an approximation. With a trace, a call's share is the time it spent in the model; requests batched together each count their full time, so this measures time on the GPU, not compute used alone.
• Idle cost needs GPU metrics to see quiet hours. Without them, only whole days with no calls count as idle. NVIDIA's utilization counts a GPU as busy whenever any work runs on it, and vLLM's signal only shows whether a request was running, so both can read higher than real use.
• A call's cost on a finished day is recorded in the audit history when it changes. During the running day the split changes every hour and isn't recorded; the call's breakdown shows its day, rate and whether the figure is still provisional.
• The move analysis compares cost only. An open model on your GPUs is not the model the API ran, its tokenizer counts the same text differently, and a pool's throughput is assumed to hold for the new traffic. It is not advice to buy GPUs or change models.

No warranty

The website is provided for informational purposes during product validation. It may change, become unavailable, or contain incomplete information while the product is being developed.