Integrations

Envoy AI Gateway is a Kubernetes-native gateway for LLM inference traffic. Waldur provisions per-customer API keys directly from marketplace orders and reports per-key token usage back for metering and billing. Keys live in Kubernetes Secrets read by the Envoy SecurityPolicy, and a two-secret pause model suspends access at the authentication layer without regenerating key material. Token and cost budgets are enforced by Waldur: usage is reported, the resource is paused when a limit is reached, and the key is blocked.

LiteLLM proxy provides a unified API across LLM providers. Waldur manages the full virtual key lifecycle - provision, rotate, pause, restore and terminate - from marketplace orders, and reports per-resource input and output tokens plus USD spend from LiteLLM’s own daily spend table. Waldur limits are mirrored onto each key as budget, TPM and RPM caps so a burst between reporting cycles cannot outrun metering, and keys can be restricted to a per-offering model allowlist.