AUDR
An open standard introduced by Chargebee for tracking AI agent operating costs. It defines a common JSON Schema that enables different frameworks, routers, and billing systems to standardize the recording and exchange of cost and attribution information for agent runs.
Timeline (last 90 days)
AUDR is a related project, with address https://openaudr.dev/.
KORA Doctor was built by a Reddit developer after the release of AUDR.
KORA Doctor's open-source code is hosted on GitHub at Krako-Labs/kora-doctor, and AUDR-related resources are at openaudr.dev.
AUDR defines a universal JSON Schema that allows any framework, router, or billing system to produce and consume such records.
AUDR relies on three core design rules: a shared run ID, clear field ownership, and strict merge rules.
In AUDR, each run has a shared run ID generated by the execution harness, which threads through all systems touched by that run.
AUDR (Agent Usage Detail Record) is an open standard introduced by Chargebee for capturing who initiated an agent run and the actual cost of that run in every system it touches.
AUDR replaces the telecom billing unit of 'call minutes' with AI-native measurement dimensions such as number of model invocations, token consumption, and tool call costs.
AUDR's positioning is similar to that of OpenTelemetry in the distributed tracing space—it does not replace individual implementations but provides a unified data contract.
The true value of an open standard depends on ecosystem adoption, and whether AUDR can become the 'CDR' of the AI agent space hinges on how many execution frameworks, routers, and billing systems are willing to integrate with it.
3 more timeline events
All Facts (13)
AUDR is a related project, with address https://openaudr.dev/.
50%UnverifiedKORA Doctor was built by a Reddit developer after the release of AUDR.
50%UnverifiedKORA Doctor's open-source code is hosted on GitHub at Krako-Labs/kora-doctor, and AUDR-related resources are at openaudr.dev.
50%UnverifiedAUDR defines a universal JSON Schema that allows any framework, router, or billing system to produce and consume such records.
50%UnverifiedAUDR relies on three core design rules: a shared run ID, clear field ownership, and strict merge rules.
50%UnverifiedIn AUDR, each run has a shared run ID generated by the execution harness, which threads through all systems touched by that run.
50%UnverifiedAUDR (Agent Usage Detail Record) is an open standard introduced by Chargebee for capturing who initiated an agent run and the actual cost of that run in every system it touches.
50%UnverifiedAUDR replaces the telecom billing unit of 'call minutes' with AI-native measurement dimensions such as number of model invocations, token consumption, and tool call costs.
50%UnverifiedAUDR's positioning is similar to that of OpenTelemetry in the distributed tracing space—it does not replace individual implementations but provides a unified data contract.
50%UnverifiedAUDR's merge rules adopt an 'append-only, no-overwrite' design: when conflicts arise, conflicting records are directly rejected, and any corrections must appear as new records.
50%UnverifiedThe AUDR project received 188 votes and 24 comments on Product Hunt, at one point ranked 2nd, and is categorized under Open Source, Developer Tools, and GitHub.
50%UnverifiedThe design philosophy of AUDR draws on the CDR (Call Detail Record) mechanism from the telecommunications industry.
50%UnverifiedThe true value of an open standard depends on ecosystem adoption, and whether AUDR can become the 'CDR' of the AI agent space hinges on how many execution frameworks, routers, and billing systems are willing to integrate with it.
50%