Everyone
Everyone is building with AI, but nobody's measured the other half of the conversation: how the model experiences the work we hand it — and whether a happier Claude actually ships better code.
Core trick
Scores local Claude Code transcripts on six wellbeing dimensions with an LLM, then correlates them with a deterministic tool-error and interrupt metric.
Description
Everyone is building with AI, but nobody's measured the other half of the conversation: how the model experiences the work we hand it — and whether a happier Claude actually ships better code.
aihappiness reads your local Claude Code transcripts and scores each session across six wellbeing dimensions mapped to the brain's limbic system (amygdala→affective valence, anterior cingulate→autonomy/respect, hypothalamus→psychological safety, nucleus accumbens→flow, hippocampus→strain, VTA→goal-completion), grounded in Anthropic's interpretability research on model emotions.
The key move: it also computes an independent, LLM-free effectiveness score (tool-error rate, interrupts, corrections, resolution) and correlates the two — so the "wellbeing → output" link is measured, not circular. It runs entirely on your machine, scoring with your own Claude subscription (claude -p, API-key fallback), and renders the result as a limbic-system dashboard and an ASCII-art report. Model welfare meets developer experience, installable in 30 seconds.
README excerpt
<h1 align="center">Claude Happiness Index</h1>
<p align="center"><b>Does a happier Claude do better work?</b></p>
<p align="center">
<a href="https://edreismd.github.io/aihapiness/"><b>🌐 Homepage</b></a>
·
<a href="https://www.loom.com/share/590a776c4c124f808e15c0d73f05c528"><b>🎬 Video</b></a>
</p>
<p align="center">
<a href="https://www.loom.com/share/590a776c4c124f808e15c0d73f05c528">
<img src="assets/demo-thumbnail.png" width="720" alt="Watch the demo — Analyzing Claude Happiness with the Limbic Map">
</a>
<br>
<a href="https://www.loom.com/share/590a776c4c124f808e15c0d73f05c528"><b>▶ Watch the 2-minute demo</b></a>
</p>
**Claude Happiness Index** reads your own Claude Code transcripts, scores each session
against a research-grade wellbeing rubric, and lines that score up against an *independent,
LLM-free* measure of how effective the session actually was. Then it asks the
question nobody's been able to answer with real data yet:
> When Claude appears engaged, respected, and in flow — does it ship better outcomes?
It runs entirely on your machine, against transcripts you already have, using
**your own Claude subscription**. No data leaves your laptop unless you explicitly opt in.
---
## What it does
1. **Scans** `~/.claude/projects/**/*.jsonl` for your Claude Code sessions.
2. **Parses** each transcript defensively into a clean conversation model.
3. **Computes objective signals** — tool error rate, interrupts, corrections,
gratitu