Earshot - From unusual sounds to repair orders
Link to open source: https://github.com/venvennnn/earshot
EarShot is a voice-first maintenance assistant that helps technicians investigate unusual machine sounds with a phone. A new rattle in a pump or a rough hum in a motor can be an early sign of trouble. EarShot helps teams capture that sound, compare it with a healthy recording, and decide whether the machine needs attention.
A technician first records the machine while it is running normally. Later, they can say, “Check Pump P‑14,” and EarShot opens the recorder. A beep signals when to begin, and a live waveform shows that the microphone is picking up sound. EarShot compares the clip with the saved baseline, shows how unusual it sounds, and can suggest a possible cause. The technician can play both recordings to hear the difference.
EarShot also supports tap testing. A technician records one tap on a known-good part, plays it back to check the reference, and then taps the parts under inspection. If a new tap sounds different or its ring dies away sooner, EarShot flags the part for a closer look.
The dashboard gives supervisors a shared view of machine checks, findings, and decisions awaiting approval. When EarShot proposes a repair, the team can review the evidence before approving a work order. With Jira Cloud connected, the approved job can also create a linked ticket containing the asset, finding, and approval details.
By keeping the recordings, findings, and repair decisions together, EarShot helps teams follow up on what technicians hear on the floor and gives the next shift a clear handover.
This build was uploaded as a hackathon project



