AUDR by Chargebee

For engineering and platform teams running AI agents in production, a single run touches multiple systems — the application knows the customer, the router knows tokens and cost, and tools know what they executed — but there is no shared way to join these into one end-to-end record. Usage data ends up orphaned from the business context that gives it meaning, so you cannot answer who ran what and what it cost. AUDR adds a shared run ID, clear field authority, and strict merge rules so every layer keeps reporting what it already reports while the records come together. For example, a platform engineer trying to bill a customer for agent usage has token counts in the router and customer IDs in the app, but no single record tying them together.