HN Showcase
Back to browse
GitHub Repository

Claude Code & Codex Session Analytics

281 starsTypeScript

We analyzed 1,573 Claude Code sessions to see how AI agents work

by keks0r·Mar 12, 2026·144 points·86 comments
Visit ProjectView on HN

AI Analysis

●●●BangerZero to OneDark Horse

First analytics layer for Claude Code revealing 26% session abandonment rate.

Strengths
  • •Genuinely novel category—no existing benchmarks for agentic session performance
  • •Real insights from 1,573 sessions like skills used in only 4% of sessions
  • •Open source with self-hosting option despite handling sensitive session data
Weaknesses
  • •Privacy concerns with uploading full session transcripts including source code
  • •Unclear value proposition for individual users versus teams
Target Audience

Teams and individuals using Claude Code for development

Post Description

We built rudel.ai after realizing we had no visibility into our own Claude Code sessions. We were using it daily but had no idea which sessions were efficient, why some got abandoned, or whether we were actually improving over time.

So we built an analytics layer for it. After connecting our own sessions, we ended up with a dataset of 1,573 real Claude Code sessions, 15M+ tokens, 270K+ interactions.

Some things we found that surprised us: - Skills were only being used in 4% of our sessions - 26% of sessions are abandoned, most within the first 60 seconds - Session success rate varies significantly by task type (documentation scores highest, refactoring lowest) - Error cascade patterns appear in the first 2 minutes and predict abandonment with reasonable accuracy - There is no meaningful benchmark for 'good' agentic session performance, we are building one.

The tool is free to use and fully open source, happy to answer questions about the data or how we built it.

Similar Projects