Principles in practice

This page summarises how the principles on the previous three pages translate into product behaviour you can observe.

Routing

Decision	What we do
Default route	None. The `model` field is required. `auto/<creator>/<model>` keeps the model your choice and ranks routes on sovereignty, then carbon, then price.
Pinning	Any caller can pin model, provider, and region per request, or per API key.
Failover	A provider outage routes to the next eligible option in the same region. We never fail across regions silently.
Carbon weight	Routing is biased toward lower-carbon regions when other constraints allow. The bias is configurable and the weight is documented in routing.

Decision	What we do
Pricing model	Pre-paid credits. The EUR price per 1M tokens for each model is shown on the model browser before you call it.
Fees	No per-token markup. Tokens are billed at the upstream provider’s price converted to EUR. Our fees apply at credit top-up and are published on the pricing page.
Free tier	None. Trying things out costs the same as production usage.
Refunds	A failed request that produced no upstream charge does not consume credits.

The full pricing rules are in credits and billing.

Decision	What we do
Prompt logging	We log token counts, model, provider, region, and timing. We do not log prompt or response content.
Retention	Per-request usage rows are kept for 90 days for billing, then rolled into daily aggregates and deleted. Aggregates are kept longer.
Export	Usage history can be exported as CSV from the dashboard.
Subprocessors	Listed on the sub-processors page, generated from the same provider catalogue the router uses.

If you need a Data Processing Agreement, contact us through the channel listed on the legal page.

Decision	What we do
Status page	Linked from the dashboard footer.
Incident communication	Public post-mortems for incidents that affected billing or routing decisions.
API stability	Breaking changes are versioned (`/v1`). Additive changes are documented in the changelog.

We do not auto-upsell to larger models. The defaults aim at the smallest model that produces an acceptable response.
We do not show comparative quality charts between providers. We are not the right venue to make those judgements; tools that benchmark outputs against reference suites are.
We do not stockpile features for the sake of feature parity. New endpoints and dashboards are added when there is a defensible reason.