A maker recently built a compact portable voice recorder designed specifically for capturing ideas before they are forgotten. Instead of relying on a smartphone app, the project focuses on a dedicated hardware workflow built around an ESP32, combining local audio recording with AI-based transcription and summarization. The device records audio directly to an SD card, provides onboard playback controls, and later converts recordings into structured text notes using cloud AI models.
What makes the project interesting is the amount of custom engineering involved despite the device appearing relatively simple externally. The recorder uses an INMP441 I2S microphone, LiPo-powered battery system, magnetic encoder feedback mechanism, and a custom interface for playback navigation. One unusual design choice is the use of a small DC motor as a tactile interaction element, allowing users to physically interrupt or control playback through motion. On the software side, recorded audio is processed using OpenAI’s Whisper model for transcription and Anthropic’s Claude API for summarization.
The most technically demanding aspect appears to be the custom double-sided 3D-printed PCB. Instead of traditional PCB fabrication, the creator manually applied adhesive copper foil onto printed traces, drilled vias by hand, and assembled a multilayer layout to fit everything inside the enclosure. The enclosure itself also includes printed live-hinge buttons, which require careful orientation relative to print layer direction for durability. Overall, the project feels less like a simple voice recorder and more like an experiment in combining embedded hardware, fabrication techniques, and AI-assisted workflows into a practical everyday tool.
