A python web app that brokers acces to certain speech-related Google AI services
- Authentication with Firebase Auth
- Enumeration of available languages
- WebSockets channels for transcribing and translating audio data
- Docker
- GCP account with Cloud Speech-to-Text API, Cloud Translation API, and Firebase Auth enabled
- A service account with read access to those APIs and a JSON credentials file
- Enable all of those APIs in your GCP project. Make sure to enable Firebase Auth in your Firebase project
- Below is a quick, safe Firebase Auth configuration for playing around with this system. All of these options are available in the setup wizard when you initially enable Auth:
- enable only email/password auth
- manually create a user
- disable user creation via the API
- Below is a quick, safe Firebase Auth configuration for playing around with this system. All of these options are available in the setup wizard when you initially enable Auth:
- Create a service account with access to these APIs. Download its credentials file and put it in the root of this project with the name
google-services.json - Build the image and run it in a container:
$ docker build -t <image_name> . && docker run -it -p 8080:8080 --name <image_name> <container_name>- Confirm its running and available (should return a list of languages):
$ curl http://<ip_appdress>:8080/getSupportedLanguages