# Get complete end-to-end agent trace visibility with MLflow

> Production agent systems demand visibility. Agno now integrates with MLflow to deliver complete, end-to-end trace observability across every model call, tool invocation, and agent step—without custom…

- Published: 2026-03-17
- Author: Agno Team
- Category: Changelog
- Canonical: https://agno-com-nine.vercel.app/articles/full-agent-trace-visibility-now-available-with-mlflow
- Markdown: https://agno-com-nine.vercel.app/articles/full-agent-trace-visibility-now-available-with-mlflow.md

Production agent systems demand visibility. Agno now integrates with MLflow to deliver complete, end-to-end trace observability across every model call, tool invocation, and agent step—without custom instrumentation or additional configuration overhead.

With a single call to `mlflow.agno.autolog()` at startup, all agent activity is automatically captured and surfaced in the MLflow UI. This applies to both individual agents and full AgentOS deployments.

Details**:**

- Full trace capture across model calls, tool use, and agent steps — out of the box
- Works with self-hosted and managed MLflow servers (AWS, Azure, GCP)
- Supports AgentOS applications with no additional setup beyond the single autolog call
- Traces are OpenTelemetry-native, making them compatible with existing observability pipelines

View the [MLflow docs](https://docs.agno.com/observability/mlflow) for more.