I built an AI Tamagotchi before Meta could
A senior editor at The Verge built a custom AI-powered gadget resembling Meta's upcoming Muse Charm for just $50, leveraging an open-source devkit and an AI coding assistant.
Intelligence analysis by Gemini 2.5 Flash
The author created a personalized AI companion, akin to a Tamagotchi, using a Waveshare ESP32-S3 1.75C pendant and Claude Code. This DIY project allowed for extensive customization, adding features like conversation history, voice output, remote camera control, and a virtual pet game, demonstrating the potential for accessible AI hardware development.
Imagine you have a tiny digital pet, like a Tamagotchi, that can talk to you and even help you take pictures! Instead of buying one from a big company, someone used a small, cheap computer and asked a super-smart computer helper, named Claude, to write all the instructions. Claude helped make the pet talk, play games, and even turn the little computer into a pocket watch, all by just being told what to do, even without knowing how to write computer code.
Analysis
The article details a compelling personal project that underscores the evolving landscape of AI hardware and software development. By constructing a functional AI companion device for a mere $50, the author effectively preempts Meta's anticipated Muse Charm, showcasing how readily available components and AI coding assistance can empower individuals to create sophisticated gadgets.
Muse Charm
Meta's vision for the Muse Charm is presented as a cute, squishy-faced AI helper designed to be carried around like a Tamagotchi. This commercial product aims to integrate AI into daily life through a portable, interactive device. The author's project serves as a direct comparison, suggesting that the core functionality and appeal of such a device can be replicated and even enhanced outside of a corporate ecosystem.
Meta's strategy involves selling these devices, but also providing open-source code and tokens for 'Muse Gadgets' to connect to its Muse AI cloud. This dual approach indicates a recognition of the DIY community's potential, even as the company plans its own proprietary hardware. The article implicitly questions the value proposition of Meta's commercial offering when a highly customizable alternative is achievable at a fraction of the cost.
Waveshare ESP32-S3 1.75C
The foundation of the author's DIY AI Tamagotchi is the Waveshare ESP32-S3 1.75C, a 1.75-inch touchscreen computer pendant. This $50 devkit, drop-shipped from China, is notably sanctioned by Meta as a suggested device for building Muse Gadgets, indicating its compatibility with Meta's open-source AI framework. Initially, the device presented limitations, such as slow responses from Muse AI, lack of audible replies, and a basic single avatar.
The initial setup involved plugging the device into a PC, pointing Claude at Meta's GitHub, and pairing it with the Muse phone app. Despite Meta's keys being installed, the AI initially misinterpreted audio inputs and lacked essential features for a seamless user experience. This highlights the gap between raw hardware capability and user-friendly functionality, a gap that the author bridged through subsequent AI-assisted coding.
Claude Code
The transformative element in this project was the application of Claude Code, an AI coding assistant. The author, possessing no prior coding experience, leveraged Claude to reprogram the Waveshare pendant, addressing its initial shortcomings. Claude facilitated the creation of a conversation history page, integrated a British AI voice from ElevenLabs for audible responses, and enabled remote camera control with live preview and a shutter button.
Furthermore, Claude was instrumental in developing a batch of diverse avatars and a Tamagotchi-like minigame with varying levels of hunger, boredom, and dirt. It also refined the push-to-talk functionality to prevent accidental audio truncation and even built a working pocket watch face and a video player for the device. This demonstrates the profound impact of AI coding tools in making complex development accessible to non-programmers, allowing for rapid iteration and feature implementation based on simple natural language requests.
Key points
- A senior editor built a custom AI-powered gadget for $50, resembling Meta's upcoming Muse Charm.
- The project utilized a Waveshare ESP32-S3 1.75C devkit, sanctioned by Meta for 'Muse Gadgets'.
- Claude Code, an AI coding assistant, was used to extensively customize the device, adding features like voice, conversation history, remote camera control, and a Tamagotchi-like game.
- The author, with no prior coding experience, demonstrated the accessibility of AI-assisted hardware development.
- The custom device offers functionalities that Meta's initial build lacked, highlighting the power of personalization.
This project demonstrates the immense potential for individuals to customize and personalize AI hardware, fostering innovation beyond corporate offerings. It suggests a future where AI companions are not just consumer products but highly adaptable tools built and tailored by users, making advanced technology more accessible and user-centric.
The reliance on Meta's servers for AI responses introduces a point of failure, as evidenced by reported lag, and raises concerns about data privacy and trust. Furthermore, the quality and security of AI-generated code remain uncertain, potentially leading to unforeseen flaws or vulnerabilities in custom-built devices.
