The Beep Locator UI displays triangulated beep positions as red Xs on a 2‑D map of the room.
AI-POWERED BEP LOCATOR PINPOINTS MYSTERIOUS BEEP TO CARBON MONOXIDE DETECTOR
*A home‑grown Python/WebSocket system triangulated an intermittent beep to within 40 cm. The source: an abandoned carbon‑monoxide alarm demanding fresh batteries.*
A persistent beep haunted a cluttered box room for weeks, its origin hidden among stacked cartons. The homeowner refused to sift through every package, so he turned to Claude, an AI code‑generation service, to build a triangulation system. Within days, a Python web app coordinated five everyday devices—two iPads, a phone, a MacBook Pro, and a Surface Ultra—into a synchronized listening array. The system captured audio, flagged beeps with FFT analysis, and used a chirp‑based alignment to locate the source. When the software reported a position, the homeowner excavated a single pile and found an ancient carbon‑monoxide detector demanding fresh batteries.
How the Beep Locator was built
Claude Code generated a Python web server that coordinates listeners via WebSockets. Five consumer devices—two iPads, a phone, a MacBook Pro, and a Surface Ultra—joined the server through a QR‑code URL. Each device recorded a 60‑second audio buffer and ran a simple FFT‑based high‑pass filter to flag potential beeps. When a beep was reported, the server instructed the device with the lowest round‑trip time to emit a short chirp. All listeners then uploaded clips from 0.6 s before to 0.8 s after the chirp, allowing precise alignment. The UI displayed icons that could be dragged in “Rough mode” to supply approximate positions, which the algorithm refined automatically.
Can consumer devices triangulate sound?
The system proved that off‑the‑shelf hardware can locate a sound source without dedicated infrastructure. By leveraging the BeepBeep paper—"A High‑Accuracy Acoustic‑Based System for Ranging and Localization Using COTS Devices" (Peng et al., 2007)—the project replicated two‑way sensing, self‑recording, and sample counting with only browsers, microphones, and speakers. The algorithm achieved a reported accuracy of about 40 cm, matching the academic benchmark. No custom firmware or external beacons were required; the entire pipeline ran in standard web browsers.
What the system discovered
After weeks of intermittent beeping, the locator placed the source within a 40 cm radius. The author dug at that spot and uncovered an ancient carbon‑monoxide detector whose battery had died. The beep frequency matched the detector’s alarm tone, confirming the system’s identification. The discovery halted the nuisance and averted a potential safety hazard, illustrating the practical value of rapid acoustic forensics in everyday environments.
Implications for DIY acoustic surveillance
The success demonstrates that hobbyists can deploy ad‑hoc acoustic networks for troubleshooting, security, or research. By reusing existing devices, the cost barrier drops to zero, and the software stack remains open‑source. The approach also validates the relevance of older academic work—BeepBeep’s 2007 methods—when paired with modern AI‑assisted code generation. As more households adopt smart speakers and IoT microphones, similar triangulation could become a standard diagnostic tool, raising both convenience and privacy considerations.
Frequently asked questions
How did the Beep Locator determine the beep's location?
It recorded audio on five devices, used FFT and a high‑pass filter to detect beeps, emitted a chirp for alignment, and triangulated the source to within about 40 cm.
What hardware was required for the system?
Two iPads, a phone, a MacBook Pro, and a Surface Ultra served as listeners; no special sensors were needed beyond built‑in microphones and speakers.
What was the actual source of the mysterious beep?
An ancient carbon monoxide detector with dead batteries, found after the system pinpointed its location.
The episode shows that AI‑driven, low‑cost acoustic tools can turn a maddening mystery into a precise fix. As more developers replicate the workflow, households may soon have instant sound‑source mapping at their fingertips—turning every beep into actionable data and reshaping how we monitor our own spaces.
Sources: [Hacker News – Locating the source of a mysterious beep (https://simonwillison.net/2026/Nov/9/mysterious-beeps/)]