Documentation

Welcome to Guardy

The AI agent observability platform that doesn't just monitor — it diagnoses and fixes issues automatically.

The AI agent observability platform that doesn't just monitor. It diagnoses and fixes issues automatically through code integration.

How It Works

Observe

Track all agent sessions, events, and user interactions in real-time.

Detect

AI automatically identifies issues: failures, frustration, confusion, and more.

Diagnose

Get AI-powered root cause analysis and recommended fixes for each issue.

Fix

Apply fixes directly to your codebase with AI assistance and create PRs.

Install the SDK, track your first agent session, and see issues detected automatically.

Install the SDK, track your first session, and detect issues in 5 minutes.

Use the Guardy SDK for rich agent tracking, or send OpenTelemetry traces from an existing pipeline.

Sessions & Events

Track every agent interaction. See LLM calls, tool usage, decisions, and user messages. Full visibility into what your agents are doing.

Automatic Issue Detection

AI scans sessions for problems: task failures, user frustration, confusion, tool errors, slow responses, and more. No manual tagging required.

AI Root Cause Analysis

Click any issue to get an AI-generated diagnosis. Understand why something is failing, not just that it's failing.

Code Integration

Connect your GitHub repository. Chat with AI about issues, get code suggestions, see diffs, and create PRs — all from one interface.

Install the SDK, track your first session, and detect issues in 5 minutes.

Learn how automatic issue detection and AI diagnosis work.

Connect GitHub, chat with AI, and fix issues directly in your codebase.

Compare native SDK tracking with OpenTelemetry trace ingestion.

Complete API reference for tracking sessions and events.

Can't find what you're looking for? Reach out to us at [email protected].