A Garmin Fenix smartwatch has been shown accepting running efficiency metrics from an external device built around ESP32 or nRF52832 hardware. The finding matters because it suggests Garmin’s sensor ecosystem is more flexible than many users assumed.
For ordinary users, the practical impact may still be limited. For makers and accessory developers, however, it points to a possible route for building Garmin-compatible hardware without relying entirely on official sensors.
What the device managed to send
In the demonstration, the Fenix recognized externally sent data as if it were native input. The metrics involved were ground contact time and vertical oscillation, two measurements often used to assess running efficiency.
That does not mean a finished DIY sensor is already ready for broad use. The project still relied on sample data rather than real readings from a fully functional custom sensor.
Who built it and how
The work was carried out by Sam Dumont, who shared the results on Reddit, in two blog posts, and on GitHub. Dumont said Claude helped with the development because he does not have deep expertise in Bluetooth Low Energy or reverse engineering.
He also said he has been familiar with Garmin’s platform and its quirks since 2020. In his view, Claude can act like a critical technical partner, although a solid grasp of the basics is still necessary.
Why the result stands out
The project is interesting not only because it worked, but because of what it reveals about the development process. It shows how AI-assisted tools can help test ideas and support unconventional software and hardware experiments.
Garmin’s ecosystem is broad, but it is not fully open. Users cannot simply pair any random sensor with a Garmin smartwatch and expect metrics such as running efficiency to appear.
That is why this demonstration has drawn attention. By combining the right chip and a careful approach, the Fenix can be made to accept external data that was previously difficult to access.
It is still unclear how widely other developers will adopt the method. Even so, the experiment suggests that unofficial Garmin accessories may have more room to grow than many expected.
