AI Observability

Use Playground and Evaluations

Test prompts, compare providers, run evaluations, and inspect model output before changing production flows.

Playground steps

  • Open AI Observability > Playground.
  • Choose provider, model, temperature, and other settings.
  • Enter a prompt and run it.
  • Compare output, latency, tokens, cost, and errors.
  • Save useful findings before changing application code.

Evaluation steps

  • Open Evaluations and choose the model and judge settings.
  • Run the evaluation.
  • Review pass rate, scores, error output, and examples that failed.
  • Use the result to update prompts, routing, or model choice.

Need more help?

Keep moving with the right next guide

If your team still runs into setup issues, empty dashboards, billing delays, callback failures, or alerting problems, continue with troubleshooting before re-running the entire onboarding flow.