Skip to main content
After you instrument your app, every LLM call sends a detailed span to your Axiom dataset. The spans are enriched with standardized gen_ai.* attributes that make your AI interactions easy to query and analyze. Key attributes include the following:

​
Span identification

  • gen_ai.capability.name: The high-level capability name you defined in withSpan.
  • gen_ai.step.name: The specific step within the capability.
  • gen_ai.operation.name: The operation type. For example: chat, execute_tool.

​
Model information

  • gen_ai.provider.name: The model provider. For example: openai, anthropic.
  • gen_ai.request.model: The model requested for the completion.
  • gen_ai.response.model: The model that actually fulfilled the request.
  • gen_ai.output.type: The output type. For example: text, json.

​
Token usage

  • gen_ai.usage.input_tokens: The number of tokens in the prompt.
  • gen_ai.usage.output_tokens: The number of tokens in the generated response.

​
Messages

  • gen_ai.input.messages: The full, rendered prompt or message history sent to the model (as a JSON string).
  • gen_ai.output.messages: The full response from the model (as a JSON string).
  • gen_ai.response.finish_reasons: The reason the model stopped generating tokens. For example: stop, tool-calls.
  • gen_ai.response.id: The unique identifier for the model response.

​
Tool attributes

  • gen_ai.tool.name: The name of the executed tool.
  • gen_ai.tool.call.arguments: The arguments passed to the tool (as a JSON string).
  • gen_ai.tool.call.result: The result returned by the tool (as a JSON string).

​
Additional attributes

For a more thorough list of attributes, see the OpenTelemetry Semantic Conventions for Generative AI.

​
What’s next?

After capturing and analyzing production telemetry: