-
Notifications
You must be signed in to change notification settings - Fork 5
[wip] realtime audio transcriptions #2092
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
|
@foolmoron I've run this with a locally running choreographer. it connects to the choregrapher which seems to return the needed key but the client throws some error. do you have everything you need to run the choreographer locally? we could also deploy the endpoint to our staging server without too much trouble. |
I did find how to get the local choreo working (there are 2 .env files that need the API URL changed) but yeah there's some client errors I haven't figured out. Do you think you could take a look and see if you can fix it faster than I could? |
|
@foolmoron Looking into it! |
|
@foolmoron Hey I just put 45 minutes into debugging. I looked for errors in the logic of the file itself and just tried commenting out various pieces to see where the issue might be coming from. It's coming from this: I've got a meeting now. I think I'd need to go deep into FlutterSoundRecorder to see if there was some error in the setup sequence. That is my suspicion. Or maybe it's incompatible with AudioRecorder. |
|
@wcjord Hm ok, might need a bigger refactor to use just one recorder |
|
@foolmoron That would probably make sense. |
…ve_splash-2.4.6 build: (deps): bump flutter_native_splash from 2.4.4 to 2.4.6
Adds
flutter_soundlibApparently
recorddoes not support audio recording streams well on web, so I pulled this one in which does. It also has a nice API that works with Dart streams.Creates websocket connection with Deepgram speech-to-text API using temp key from Choreo
See: https://github.com/pangeachat/2-step-choreographer/pull/490
Adds text widget to recording dialog that uses StreamBuilder
This uses the websocket connection and builds a string using the text responses from the Deepgram server (not well tested right now)