Voice Inspection App
Log inspections by speaking. The Whisper speech model runs on your phone — not on a server — so transcription works in a dead-zone out-yard and your voice never goes to a speech server.
Features
Hands‑free logging
Speak your inspection notes; the app transcribes them instantly, even with gloves.
On-device Whisper (~181 MB)
The speech model is stored locally after a one-time Wi-Fi download; transcription is local compute, not a cloud call.
Full integration
All notes sync with HiveSense inspections when you get signal.
Why the model lives on your phone, not the cloud
Most voice features in beekeeping apps work by recording your audio, uploading it, and waiting for a server to send text back — no signal, no notes. HiveSense downloads the ~181 MB Whisper model once over Wi-Fi (whisper.rn), so transcription runs on your phone's own processor: it works with airplane mode on, and nothing you say is sent to a speech service. It's the same reason the rest of HiveSense runs on an on-device database (WatermelonDB) — inspections happen in the field, and remote out-yards are exactly the places where the cloud isn't an option.
Read next
FAQ
Does voice inspection work offline?
Yes. The Whisper model runs entirely on your device, so you can dictate notes without any cell signal.
Do I need an internet connection to download the voice model?
You need Wi‑Fi once to download the ~181 MB model. After that, all processing is local.
Why does the Whisper model live on your phone instead of the cloud?
Most voice features record your audio, upload it, and wait for a server to send text back — no signal, no notes. HiveSense downloads the ~181 MB Whisper model once over Wi-Fi (whisper.rn), so transcription then runs on the phone itself: it works with airplane mode on, and nothing you say is sent to a speech service. It is the same reason the rest of HiveSense runs on an on-device database — inspections happen in the field, and remote out-yards are exactly where the cloud is not an option.