obody actually enjoys talking to an artificial voice that babbles endlessly when a five-second glance would do. Today's commercial alternatives force you into a grim compromise: you either type prompts on a keyboard like it is 1995 or endure conversational AI on a glowing phone screen while dodging push notifications. Independent maker JeanDeBaill bypassed that corporate design trap entirely by building the Claude Pager—a pocket-sized DIY terminal dedicated strictly to querying Claude and Grok without the digital noise.
The concept cuts through modern smartphone clutter by splitting the interaction cleanly down the middle: you speak your prompt out loud, and the answer appears silently as text on an electronic paper screen. As Cameron Coward highlighted for Hackster.io, this setup eliminates the friction of sitting through synthesized voice responses while keeping physical controls down to basic clickers.
This pager lets him speak prompts out loud and see the response on the E Ink screen.
Instead of juggling bloated apps, glaring OLED glass, and fragile battery meters, you press a tactile button, talk, and read the answer on matte electronic paper. It is a direct rebuke to commercial misfires like the Rabbit R1 and Humane Ai Pin, which overcomplicated the exact same premise.
Hardware Stripped to Essentials
Under the hood, JeanDeBaill built the Claude Pager around Heltec’s Vision Master E290 development board, which pairs an ESP32-S3 microcontroller and onboard lithium battery management with a low-power 2.9-inch E Ink display running at a 296 × 128 resolution. The entire assembly snaps cleanly into a custom 3D-printed enclosure shaped like a retro pocket pager.
On the software side, JeanDeBaill routes speech processing through Whisper for Speech-to-Text via Groq, paired directly with Claude Haiku 4.5 through its native API. Because electronic paper sips power only when redrawing pixels, the device achieves practical multi-day endurance that standard mobile screens cannot touch. While Big Tech spent millions over-engineering fragile AI wearables that nobody asked for, an independent workshop proved that useful dedicated hardware only takes an ESP32-S3, a 2.9-inch e-paper panel, and deliberate restraint.