Text to speech REST API for multiple TTS engines.

You can send a text to be converted into audio, using different TTS engines and sound effects. Then, you will be listening to it on your local audio device, or receiving the resultant audio file.

First, you should install the supported TTS engines:

apt install python3 sox libsox-fmt-mp3
pip install google_speech

apt install python3 sox libsox-fmt-mp3
pip install gTTS

apt install festival festvox-ellpc11k

apt install espeak

You also need to install nodejs and npm, and then, simply run npm install and npm start. The API should now be running at http://localhost:3000.

Or you can just use pedroetb/tts-api Docker image, which already has all dependencies configured.

The only requirement is to have Docker installed. Then, you can run:

docker run --rm -d --name tts-api --device /dev/snd -p 3000:3000 pedroetb/tts-api

The API will be running and accessible at http://localhost:3000.

Alternatively, you can deploy it in a Docker Swarm cluster using docker compose (already included in Docker if using modern version) and docker swarm (create Swarm cluster first):

cd deploy

# Deploy Caddy service
env $(grep -v '^[#| ]' .env | xargs) \
 TRAEFIK_DOMAIN=change.me \
 docker stack deploy \
 -c compose.caddy.yaml \
 tts-api

# Run TTS-API container
docker compose \
 -f compose.tts-api.yaml \
 -p tts-api \
 up -d

The service is prepared to be reverse-proxied with Traefik, and accessible at tts.${TRAEFIK_DOMAIN} domain. How to run Traefik is not described here, check its official site.

The proxy needs a little help from Caddy, because Docker Swarm is not compatible with devices configuration (required to use sound capabilities) and Traefik cannot work with Docker containers and Docker Swarm services all at once. This way, only Caddy service is exposed using Traefik and tts-api container is only accessible through reverse-proxy provided by Caddy (same way Traefik is reverse-proxing to Caddy).

Both, Docker container and service, can be running on different hosts, because they are able to communicate through a Docker overlay network. Run tts-api Docker container on host which has speakers, so you can listen speech.

Don't forget to edit TRAEFIK_DOMAIN environment variable before deploying.

When running, the API will receive POST requests at http://localhost:3000. You can use your favourite REST client to send a request, or use the built-in form.

Both modes (playing or downloading audio) are available using different voice codes, select one according to your needs.

Go to http://localhost:3000 with your browser, fill the form with data and submit it. Just that.

You can send a POST request to http://localhost:3000 following this scheme:

• Headers
  • Content-Type: application/json
• Body
  • { "voice": "google_speech", "textToSpeech": "hello world", "language": "en", "speed": "1" }

For example, using curl:

# Play audio
curl http://localhost:3000 \
 -d '{ "voice": "google_speech", "textToSpeech": "hello world", "language": "en", "speed": "1" }' \
 -H 'Content-Type: application/json'

# Download audio file
curl http://localhost:3000 \
 -d '{ "voice": "gtts_file", "textToSpeech": "hello world", "language": "en", "speed": "1" }' \
 -H 'Content-Type: application/json' \
 -o 'output.mp3'

Google Speech is a simple multiplatform command line tool to read text using Google Translate TTS (Text To Speech) API.

You need to be online to communicate with Google servers.

Learn more at https://github.com/desbma/GoogleSpeech

Google Text-to-Speech (gTTS) is a Python library and CLI tool to interface with Google Translate's text-to-speech API.

You need to be online to communicate with Google servers.

Learn more at https://github.com/pndurette/gTTS

Festival is a free software multi-lingual speech synthesis workbench that runs on multiple-platforms offering black box text to speech, as well as an open architecture for research in speech synthesis.

It works offline.

Learn more at http://www.cstr.ed.ac.uk/projects/festival/ and http://festvox.org/festival/

eSpeak is a compact open source software speech synthesizer for English and other languages, for Linux and Windows.

It works offline.

Learn more at http://espeak.sourceforge.net/

This project is released under the MIT License.