Show HN: AgentLens – Replay and compare Codex runs

Category: devtools

Tags: ai-agents, debugging, replay, codex, developer-tools

Score: 7.0/10 (Innovation: 6, Technical: 7, Documentation: 8, Utility: 7)

AgentLens is a replay and comparison tool for OpenAI's Codex CLI that records runs, produces a local interactive timeline of commands, file changes, and token usage, and can export a redacted standalone HTML report. It also offers side-by-side comparison of two runs to highlight divergences, making it useful for debugging and sharing AI agent behavior.

Target audience: backend developers, devops, AI engineers

Repository: https://github.com/fang520huang-lgtm/AgentLens · JavaScript · MIT · 3 stars

View on Hacker News