AgenticLens

AgenticLens processes JSONL log files generated by AI agents and presents the execution data in multiple visual formats. It offers a flow view that maps the sequence of actions, a tree view that shows hierarchical relationships, and a timeline view that displays temporal ordering. The replay view lets users step through the agent’s behavior, while the inspector view provides detailed inspection of individual log entries.

The tool is aimed at developers who build and maintain AI agents and need to understand how those agents make decisions during runtime. By turning raw log text into interactive diagrams and playback controls, it helps locate logical errors, performance bottlenecks, or unexpected branching in the agent’s workflow.

AgenticLens is positioned as an experimental utility for visual debugging, tracing, and replay of agent processes, focusing on clarity and interactivity rather than code execution or deployment.