TraceLab
What it is
TraceLab is an open-source toolkit and dataset hub for analyzing the execution traces of coding agents (like Claude Code and Codex). It provides a pipeline to collect, sanitize, and visualize LLM-invocation logs, including tool calls and prompt-cache splits.
The problem it solves
The lack of structured, sanitized, and easily analyzable data regarding the internal 'thought' processes and tool-use patterns of coding agents.
What you could use it for
- Analyzing real-world coding agent workloads and performance
- Sanitizing and anonymizing private LLM interaction logs for research
- Visualizing and auditing tool-use patterns in AI agents
- Reproducing research experiments on LLM-generated code traces
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.