Trails is a platform that analyzes LLM agent execution traces to provide automated insights. It helps developers understand what went wrong during agent runs and enables them to create issues and fixes immediately.

The product provides aggregate analysis that shows patterns across all your runs, allowing you to see what's failing at a glance without reading traces one by one. It automatically detects, tags, and groups issues so you can filter and prioritize problems effectively. The system includes step-by-step replay functionality that lets you walk through each action with screenshots, browser state, and the agent's reasoning.

Trails works by having users upload their agent's execution logs, which it parses browser-use format out of the box. The system then analyzes these traces to automatically detect issues, tag them, and group them so patterns emerge across your runs.

The main benefit is that you can fix problems faster by focusing on what matters most. You can filter by issue type, drill into specific traces, and ship improvements more efficiently.

This product targets developers working with LLM agents who need to debug and improve their agent performance through detailed trace analysis.