Realtime Voice to Text from a radio

Realtime Voice to Text Project

I am happy create the project in stages.

Stage 1
Analogue audio into a raspberry Pi, audio stream sent to voice to text API
Audio online with text displayed with time stamp
Editable keyword alert, SMS, email or both
Viewable on mobile and desktop
secure login, multipule users

Stage 2
Audio from Uniden SDS200 radio scanner ethernet port
Metadata displayed

Stage 3
SDR running Raspberry Pi used instead of analogue and SDS200


The ultimate aim:
1. Send a live stream to ‘voice to text’ API, 24/7, it is not continuous talk but high load
2. Audio streaming online with test displayed.
3. Have the text available to view online
4. Audio replay matched to text if selected
5. Keywords to trigger alerts, SMS and/or email, editable
6. Multiple streams from different locations, could appear in different columns or colours
7. Each stream is a different log
8. Time stamped
9. Metadata with text
10. Hold audio files for 24 hours, hold text log for 7 days
11. Secure login for multiple users with different streams available
12. Raspberry Pi’s at each remote location to feed the audio to the voice to text API

Hosting:
?

Notes:
I understand paying for fast and accurate voice to text maybe required


Would like:
Full access to code
Operate on Raspberry Pi
Take audio from analogue audio input and/or Uniden SDS200 ethernet port with metadata about channel info, eventually using a SDR as the source.