Arize AX

Send traces to Arize AX

Arize AX is an evaluation and observability platform developed by Arize; it offers tools for agent tracing, evals, prompt optimization, and more.

Step 1: Get your Arize AX credentials

In Arize AX, navigate to your space settings to find your API key and space key:

  1. Log in to your Arize AX account
  2. Go to Space Settings to find your Space Key
  3. Go to API Keys to create or copy your API key
  4. Note the Project Name you want to use for organizing traces

Step 2: Enable Broadcast in OpenRouter

Go to Settings > Observability and toggle Enable Broadcast.

Enable Broadcast

Step 3: Configure Arize AX

Click the edit icon next to Arize AX and enter:

  • Api Key: Your Arize AX API key
  • Space Key: Your Arize AX space key
  • Project Name: The name of the tracing project in Arize AX
  • Base Url (optional): Default is https://otlp.arize.com

Step 4: Test and save

Click Test Connection to verify the setup. The configuration only saves if the test passes.

Step 5: Send a test trace

Make an API request through OpenRouter and view the trace in your Arize AX dashboard under the specified project.

Arize Trace View

Custom Metadata

Arize AX uses the OpenInference semantic convention for tracing. Custom metadata from the trace field is sent as span attributes in the OTLP payload.

Supported Metadata Keys

KeyArize MappingDescription
trace_idTrace IDGroup multiple requests into a single trace
trace_nameSpan NameCustom name for the root trace
span_nameSpan NameName for intermediate spans in the hierarchy
generation_nameSpan NameName for the LLM generation span
parent_span_idParent Span IDLink to an existing span in your trace hierarchy

Example

1{
2 "model": "openai/gpt-4o",
3 "messages": [{ "role": "user", "content": "Classify this text..." }],
4 "user": "user_12345",
5 "session_id": "session_abc",
6 "trace": {
7 "trace_id": "classification_pipeline_001",
8 "trace_name": "Text Classification",
9 "generation_name": "Classify Sentiment",
10 "dataset": "customer_feedback",
11 "experiment_id": "exp_v3"
12 }
13}

Additional Context

  • Custom metadata keys from trace are included as span attributes under the metadata.* namespace
  • The user field maps to user identification in span attributes
  • The session_id field maps to session tracking in span attributes
  • Token usage, costs, and model parameters are automatically included as OpenInference-compatible attributes

Privacy Mode

When Privacy Mode is enabled for this destination, prompt and completion content is excluded from traces. All other trace data — token usage, costs, timing, model information, and custom metadata — is still sent normally. See Privacy Mode for details.