How can I enable streaming text-to-speech in my self-hosted deployment?

Deepgram's Text-to-Speech WebSocket interface provides a real-time, interactive solution for converting streaming text to speech. This can be enabled in your self-hosted deployment by enabling a feature flag.

If you already have Aura models exposed in your deployment, these will work for both batch and streaming TTS. No additional models are needed to enable streaming TTS specifically.

Updating your self-hosted deployment

1. If you don't already have an Aura model available, request one from your Deepgram Account Representative. Place it in your models directory, accessible to your Deepgram Engine containers.

2. Update your configuration based on your container orchestrator.

   1. Docker/Podman

      1. Update your api.toml configuration file to enable the streaming TTS feature flag, as seen in the api.toml Configuration Update code block below.

      2. Update your Compose file to use a supported release. Update the image tag for API and Engine containers to release-241024 or later.

      3. Restart the API and Engine containers.

   2. Kubernetes

      1. Using a v0.7.0+ release of the deepgram-self-hosted Helm chart, set api.features.streamingTTS to true in your values.yaml file.

      2. If you have overridden the default values for api.image.tag or engine.image.tag, make sure to use release-241024 or later.

      3. Upgrade your Helm installation.

   3. To verify your setup, open a streaming TTS connection to your self-hosted deployment.

api.toml Configuration Update

...
[features]
speak_streaming = true
...

Troubleshooting

Feature flag not enabled

If you receive a 405 Method Not Allowed response, the streaming TTS feature flag may not be enabled for your Deepgram API container, or you may be on an older release that does not support the feature.

Check that your api.toml configuration file has been updated, as described in Updating your self-hosted deployment.