FreeVisit Website
Tracked since2026
The Bottom Line
- Price
- Free, no paid tier. Compare plans
Key facts
- Open standard for recording agent run attribution and usage across systems.
- Provides shared run IDs, clear field authority, and strict merge rules.
- Includes adapters for popular runtimes.
Pros
- Open and neutral standard, not tied to a single vendor.
- Works with existing runtimes and destinations without replacing them.
- Provides end-to-end visibility into agent economics.
Cons
- Requires integration effort to adopt adapters or write custom sinks.
- Still early-stage; ecosystem and tooling are evolving.
What is AUDR by Chargebee?
AUDR (Agent Usage Detail Record) is an open standard under Apache 2.0 license for tracking the end-to-end usage and cost of AI agent runs across multiple systems. Inspired by telecom's Call Detail Record, it provides a common format for capturing attribution (customer, environment, initiator) and usage (tokens, tool calls, compute time) in a single record. The standard defines three rules: a shared run ID minted by the harness, clear authority per field (e.g., harness owns attribution, router owns usage), and strict merge rules to prevent conflicts. It includes adapters for popular runtimes like Vercel AI SDK, LiteLLM, NVIDIA NeMo Relay, and Merge Gateway, as well as sinks for local files or other destinations. AUDR is stewarded by Chargebee and aims to be community-owned, with the goal of moving cost governance to an independent foundation.
Key Features
- Shared run ID minted by harness and passed through all systems.
- Clear authority per field: harness owns attribution, router owns usage.
- Strict merge rules: no component rewrites another's block; corrections are new records.
- Adapters for Vercel AI SDK, LiteLLM, NVIDIA NeMo Relay, and Merge Gateway.
- Core SDK in Python and TypeScript for custom integrations.
Pricing
Free
AUDR by Chargebee is completely free to use with no hidden costs.
Reviews
$99Free with your review
Write a reviewReview AUDR by Chargebee, get a free AI guide
Share your experience and we will send you Improve Your Thinking Patterns Using ChatGPT, free.
Best AUDR by Chargebee Alternatives
Top alternatives based on features, pricing, and user needs.
Observe.aiPaid
AI Agents for customer experience: Automate interactions, augment agents, and analyze conversations.
Weights & BiasesFreemium
Track, compare, and share ML experiments and models
MLflowFree
Manage your ML lifecycle: track, register, and deploy models
Comet MLPaid
Machine learning experiment tracking platform
ClearMLFreemium
Open-source MLOps platform for experiment tracking
Neptune.aiFreemium
Experiment tracking for ML teams
Still deciding?
Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.
Explore More
AUDR by Chargebee FAQ
How does AUDR help a developer track the cost of an AI agent run across multiple services?
AUDR (Agent Usage Detail Record) provides a common open standard for capturing attribution and usage data in a single record, including fields for tokens, tool calls, compute time, and the customer, environment, or initiator. This allows developers to log the end-to-end cost of a single agent run even when it spans multiple systems or models, without replacing their existing runtimes or destinations.
Can AUDR be used with a specific competitor like Gentrace?
AUDR is designed as an open standard under Apache 2.0 that works alongside existing tools rather than replacing them. Unlike proprietary solutions such as Gentrace, which lock usage data into a single vendor, AUDR allows teams to export records to any destination and use their choice of analytics or cost-management tools.
Which teams benefit most from adopting AUDR?
Teams building multi-model AI agents or operating across multiple cloud environments benefit most from AUDR because it provides a unified format for tracking usage and cost across those systems. It is especially useful for developer operations, FinOps, and platform engineering teams that need granular cost attribution per customer or environment.
What are the main limitations or trade-offs of using AUDR in production?
AUDR requires integration effort to adopt its adapters or write custom sinks, and it is still early-stage, meaning the ecosystem and tooling around it are evolving. Teams must be prepared to invest in setup and stay current with updates as the standard matures.
How is AUDR by Chargebee priced?
AUDR is free to use with no paid plan required, as it is an open standard released under the Apache 2.0 license. There is no cost to adopt the specification, implement adapters, or integrate with existing runtimes and sinks.
Does AUDR integrate with the Vercel AI SDK or LiteLLM?
Yes, AUDR includes pre-built adapters for popular runtimes such as the Vercel AI SDK, LiteLLM, NVIDIA NeMo Relay, and Merge Gateway. These adapters let teams emit Agent Usage Detail Records directly into the standard format without modifying their existing agent orchestration code.
Why does AUDR define a shared run ID and strict merge rules for attribution and usage fields?
AUDR defines a shared run ID minted by the harness and strict merge rules so that multiple systems can contribute attribution fields (like customer and environment) and usage fields (like tokens and compute time) into one record without conflicts. This prevents double-counting or data collision when an agent run spans a routing layer, a model provider, and a tool service.
Source: openaudr.dev