Skip to main content
Meliai takes a transparent, per-request approach to billing: every API response includes a billing_cost field that tells you exactly what was charged for that call — no aggregated end-of-month surprises. Pricing is either energy-based (measured in kilowatt-hours) or credit-based depending on the model, and you can always see which mode applied by inspecting the paid_with field.

Billing modes

Meliai supports two billing modes:

Energy billing

Charges are calculated from the actual energy consumed by the inference workload, measured in kWh. This maps directly to the environment_impact.energy_kwh value in the response and is the default for most text-generation models.

Credit billing

Some models are billed in Meliai credits — a fixed unit that abstracts over provider-specific pricing. Credits are pre-purchased and drawn down per request.
Every response body includes a billing_cost object so you always know what was charged:

Routing and cost

Your choice of routing flavor directly affects how much each request costs: Append the appropriate suffix to your model ID to activate a flavor:
Use :batch routing with the Batches API for high-volume offline workloads such as document processing or dataset annotation. Batch jobs are queued and processed during low-demand windows, giving you the lowest per-token cost available on the platform.

Free trial

You can start using Meliai without a credit card. New accounts receive a free usage allocation so you can explore the API and evaluate models before committing to a paid plan.
View current per-model pricing, top up your energy balance or credits, and manage your billing plan at https://meliai.ai/pricing.