ttsapi

A caching text-to-speech server using Google Cloud text-to-speech API

hcvst at eybuntu in ~/dev/python/ttsapi/ttsapi (master●●●) (ttsapi-EFmfwQLo) 
$ GOOGLE_APPLICATION_CREDENTIALS=/home/hcvst/dev/google-service-account-text2speech.token AUDIO_DIR=/tmp ROOT_URL=http://127.0.0.1:8000 AUTH_TOKEN=123 gunicorn server:app
[2020-06-18 22:34:34 +0200] [4481] [INFO] Starting gunicorn 20.0.4
[2020-06-18 22:34:34 +0200] [4481] [INFO] Listening at: http://127.0.0.1:8000 (4481)
[2020-06-18 22:34:34 +0200] [4481] [INFO] Using worker: sync
[2020-06-18 22:34:34 +0200] [4484] [INFO] Booting worker with pid: 4484

Docker

Build

docker build . -t tts

Run

docker run -d --name tts -v tts:/tmp/tts -e AUDIO_DIR=/tmp/tts -e EXTERNAL_URL=https://tts.bot.co.za -e AUTH_TOKEN=12345 -v /home/bot/google-service-account-text2speech.token:/tmp/google.token -e VIRTUAL_HOST=tts.bot.co.za -e GOOGLE_APPLICATION_CREDENTIALS=/tmp/google.token tts

Environment variables

AUDIO_DIR - absolute path of where to save audio files to
AUTH_TOKEN - Token clients need send in Authentication header
EXTERNAL_URL - server's address e.g. http://example.com:1234
GOOGLE_APPLICATION_CREDENTIALS - Google credentials json file

nginx-proxy (optional)

VIRTUAL_HOST - see https://github.com/nginx-proxy/nginx-proxy

Google credentials

Please consult https://cloud.google.com/text-to-speech/docs/reference/libraries#setting_up_authentication.

LD_LIBRARY_PATH

If you encounter the error ImportError: libstdc++.so.6: cannot open shared object file please point LD_LIBRARY path at the directory containing the shared library e.g.:

LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/

hcvst/ttsapi