One governed API between your company and every AI model: routing, receipts, per-request residency selection, personal-data masking and cost control — applied per request, ready for the EU AI Act era.
Which model, why, what it cost, where the data went — logged to a named person, exportable as CSV or JSON. Certificates prove the platform; receipts prove your decision.
Names, phone numbers and IDs can be detected and replaced with opaque tokens before a request leaves the Gateway — and put back in the answer before it reaches you, so the provider works on the masked text. This protection is rolling out.
Auto picks the best model your policy allows — never the cheapest — with automatic failover when a provider stumbles. Prepaid balance, per-project caps and rate limits mean requests pause at the cap instead of surprising you on an invoice.
Three residency tiers — EU-Strict, EU-Cloud, and an opt-in Global — applied on each request per the tier you choose, not promised in a policy PDF. Project-scoped API keys, created on demand and revocable one by one, keep integrations inside the same boundaries.
Two-line integration. Governance from the first request.
Xalerate