I am new to OpenHab. I would like to use my iPhone as a microphone to input voice into OpenHab, then run STT, AI and TTS models on my local GPUs. The TTS audio would be sent back to the iPhone.
From some Web research, I was assuming the OpenHab iOS app supported streaming speech, but I can’t see a microphone icon anywhere. I can’t find a widget that would add microphone support either.
Separately, I installed Vosk in OpenHab, but I could not find any instructions what to do next to use STT. I’m also unclear if/how I can get Vosk to use my GPU.
First, looking at the readme for the iOS app I see no mention of support for voice commands in the iOS app. I don’t know if that’s supported.
Assuming it’s there and just not documented and it works similar to the Android app, you’ll only see the microphone icon when displaying a sitemap. It’s not there when using MainUI. The apps natively render sitemaps but provide just a webview for MainUI so there’s no easy way to add stuff that isn’t already there.
Beyond that, all the rest of the generic voice configuration takes place under MainUI → Settings → Voice. Here is where you choose the voice interpreter and other configuration options. Instructions can be found at Multimedia | openHAB. Any settings specific to Vosk will be found under MainUI → Settings → Vosk Speech to Text (third column)
I can find no mention of GPU acceleration for Vosk so I suspect that’s not supported.
That will give you your STT but you’ll still have to select an NLP processor. If you choose Rule-Based you’ll configure a rule to receive all the text and parse and execute actions based on the commands.
