Observability that's already on

OpenTelemetry tracing, structured logging, and session-based debugging are automatic. See prompts, responses, token costs, and timing for every span. Evals show up in your traces too.

Zero setup, full visibility

Agentuity handles almost all instrumentation for you. LLM calls, API calls, storage operations, and evals are all traced automatically.

Automatic Instrumentation

Making LLM calls, API calls, or storage operations? Agentuity tracks all of it. You can add custom spans when needed, but the default coverage is comprehensive.

Session-Based View

See waterfall spans and traces grouped by session. Click into any span to see timing, prompts, responses, and costs. Logs and evals are attached to the same timeline.

Evals in Your Traces

When evals run, they appear as spans in your trace. You can see exactly where an eval ran and whether it passed or failed, right alongside your agent's execution.

Console and CLI, same data

View sessions, logs, and traces in the web console. Or use the CLI to query programmatically. On most console screens there's a terminal icon showing the equivalent CLI command. You can even SSH into deployed containers for deep debugging.

See it in action

Watch this 1-minute overview to see how to monitor and debug your AI agents using Agentuity's comprehensive observability features and real-time insights.

Resources

Logging

Structured logging with log levels, child loggers, and searchable metadata. All logs are automatically tied to session IDs for debugging.
Read

Tracing

OpenTelemetry spans for performance debugging with attributes, events, and nested spans to track timing across your system.
Read

Sessions & Debugging

Debug agents using session IDs, CLI commands, and trace timelines. Understand thread vs session scope and use metadata for filtering.
Read

Telemetry API

Reference for ctx.tracer, span attributes, events, and OpenTelemetry integration.
View

Frequently Asked Questions

  • Do I need to add instrumentation code?
  • Can I see costs per agent call?
  • How are traces organized?

Ready for complete visibility?

Monitor, debug, and optimize your agents with complete visibility.