Registration is open

  1. Arize AX
  2. Phoenix OSS
  3. Pricing
  4. Docs
  5. Learn
  6. Company
Book a demo Sign in Sign up

Subscribe to the Arize blog

Get the latest

On this page

    Suggested reading

    Bar chart titled "Correctness" comparing four arms on a 0 to 1.000 scale. LobeHub 0.826, Vault 0.833, MCP 0.834, Baseline 0.845. All four bars sit in a tight cluster just above 0.825.
    MCP vs. CLI Skills for agents: what our eval found (and which you should use)
    Diagram comparing agent behavior without and with observability. On the left, a dense, unstructured JSON trace shows raw steps like prompts, tool calls, and model responses. On the right, the same process is visualized as a clear, step-by-step timeline: user prompt, retrieval query, model reasoning, and final agent response. The image illustrates how observability turns opaque agent logs into a structured, understandable workflow.
    Why agent telemetry needs standards

    ML Observability Workshop Series

    Published Jul 28, 2025

    Events

    Views Navigation

    Event Views Navigation


    Today





    Share

    Suggested reading

    Bar chart titled "Correctness" comparing four arms on a 0 to 1.000 scale. LobeHub 0.826, Vault 0.833, MCP 0.834, Baseline 0.845. All four bars sit in a tight cluster just above 0.825.
    MCP vs. CLI Skills for agents: what our eval found (and which you should use)
    Diagram comparing agent behavior without and with observability. On the left, a dense, unstructured JSON trace shows raw steps like prompts, tool calls, and model responses. On the right, the same process is visualized as a clear, step-by-step timeline: user prompt, retrieval query, model reasoning, and final agent response. The image illustrates how observability turns opaque agent logs into a structured, understandable workflow.
    Why agent telemetry needs standards