An Android keyboard that performs speech-to-text (STT/ASR) with OpenAI Whisper and input the recognized text; Supports English, Chinese, Japanese, etc. and even mixed languages.
35
stars
4
forks
source link
Create a custom backend server for performing speech-to-text #3
After implementing this issue, the speech input should be able to perform speech-to-text by sending request to our custom server (instead of sending to the OpenAI API). The performance of the server inference speed need not be optimized in this issue.
Copy the files from ahmetoner/whisper-asr-webservice to /server, and record the commit ID in the commit description (for the ease of later patching).
Also get familiar with interacting with the provided server.
Add an option in the settings page of the speech input android app, allowing users to switch between OpenAI API (with the API key) or a custom server (with the IP and port information).
Depends on #2.
After implementing this issue, the speech input should be able to perform speech-to-text by sending request to our custom server (instead of sending to the OpenAI API). The performance of the server inference speed need not be optimized in this issue.
/server
, and record the commit ID in the commit description (for the ease of later patching).