olakai-add-monitoring

To add monitoring to an existing AI agent without changing its logic, you can wrap your OpenAI or Anthropic clients. This automatically tracks calls, tokens, and context for observability with minimal code changes.

You can track KPIs and telemetry for a deployed LLM application by configuring CustomDataConfigs. This derives actionable numeric metrics from your tracked calls and context to build monitoring dashboards.

Yes, you can monitor multi-step workflows using non-wrapped providers through manual event logging. This allows you to aggregate data from various steps into a single KPI for comprehensive observability.

To set up observability for your AI agent, you need the Olakai CLI installed and your agent API keys. You must also have the ability to wrap clients or log manual events to produce KPI data.

Yes, Olakai monitoring works with both code-instrumented apps and deployed LLM services. It automates adding observability and KPI tracking to any ongoing AI agent deployment requiring telemetry.