A open-source, local trace viewer for Claude Code and Codex sessions

Name: A open-source, local trace viewer for Claude Code and Codex sessions
Availability: InStock
Author: MediumD

by MediumD·May 13, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemNiche Gem

Chrome DevTools for Claude Code sessions when LangSmith drops local tool calls.

Strengths

•Purpose-built for agent harnesses, not generic LLM observability
•Fully local with zero setup beyond node, no cloud dependency
•Shows per-turn token splits and subagent context that generic tools ignore

Weaknesses

•Very narrow audience limited to Claude Code and Codex users specifically
•Zero GitHub stars suggests extremely early stage with unproven stability

Post Description

I built this because I had a hard time investigating why certain Claude Code subagent/skills were performing poorly. Regular old, boring software has had profilers and debuggers for decades. When a request is slow or a function blows up memory, you have tools that tell you exactly where.

In the new world order of agentic development, poor performance is usually due to context bloat - either too many unnecessary tokens or missing relevant ones. I wrote this to help answer questions like:

- Which tool calls ballooned the context window, and on which turn? - What context was the subagent/skill passed when it was spawned? - What steps were repeated every session that could be lifted into a skill/documentation or cached lookup?

Agent Profiler is the tool for just that. It requires zero setup (just node), is fully local, open-source and can support any agent harness with a local transcript format.