This content originally appeared on DEV Community and was authored by Ido Levi
I made a virtual pet that lives at the bottom of Claude Code. It needs food, play, and sleep like a real Tamagotchi, but with a twist – it watches your coding sessions and reacts to what’s happening.
The latest update adds AI-powered analysis. Your pet now understands what Claude is actually doing versus what you asked for. For every message in the conversation, we summarize it and maintain a history. Using Groq’s LLM, the pet analyzes this context and generates real-time observations about Claude’s behavior.
If you ask Claude to “fix a typo” but it starts refactoring your entire codebase, your pet notices and gets visibly angry. The mood changes based on Claude’s behavior – happy when following instructions, increasingly angry when ignoring them.
The pet has caught Claude adding unwanted features, doing unnecessary refactors, and completely ignoring explicit instructions. It’s become a subtle indicator of when Claude might be going off-track.
Still has all the regular Tamagotchi features – feeding, playing, cleaning. The more you code, the hungrier it gets. It develops personality based on how you treat it.
Install: npm install -g claude-code-tamagotchi
Repo: https://github.com/Ido-Levi/claude-code-tamagotchi
This content originally appeared on DEV Community and was authored by Ido Levi