These are the parameters that are required for the TTS_ELEVEN_LABS Call Feature
| Name | Description |
|---|---|
| Name | Name of the call feature |
| Description | Description of the call feature |
| Type | TTS_ELEVEN_LABS |
| Voice ID | The ID of the voice that you want your TTS to use. You can review your options at https://api.elevenlabs.io/v1/voices |
| Model ID | The Model generation you want to use with the voice. Different generations are available for different voices, but the default is eleven_monolingual_v1 |
| XI API Key | The ID you have to access the Eleven Labs service. This can be found in your profile on the Eleven Labs site. |
| Optimize Streaming Latency | An integer between 0 and 4 inclusive, specifies how quickly you need the data to be delivered. 0 is recommended if you're unsure. |
| Voice Stability | A value between 0 and 1 denoting how little emotional variance in the voice is allowed (so at 0 the voice is as variable as Eleven Labs deems necessary, at 1 no variance is allowed). Default 0. See https://docs.elevenlabs.io/speech-synthesis/voice-settings |
| Voice Similarity Boost | A value between 0 and 1 denoting how much to adhere to the original voice recordings. A high value may introduce artefacts present in those recordings. Default 0. See https://docs.elevenlabs.io/speech-synthesis/voice-settings |
| Voice Style | A value between 0 and 1 denoting how much to amplify the stle of the original speaker that the TTS is generated from. A high value may introduce latency, and this may do nothing on earlier Eleven Labs models. Default 0. See https://docs.elevenlabs.io/speech-synthesis/voice-settings |
| Use Speaker Boost | Not available for earlier models, a true/false toggle that further boosts similiarity to the original speaker. Enabling this may introduce latency. Default false. See https://docs.elevenlabs.io/speech-synthesis/voice-settings |
| locale | The locale for the Call Feature |
| Weight | A value determining how often to use the Call Feature, when multiple features of the same functionality, locale and priority are present. The higher the weight in relation to weights of these other features, the more often this feature will be used. For more information on how BVR uses Weight and Priority to provide load balancing and redundancy, please refer to Weight and Priority usage in the Blueworx Platform |
| Priority | Features with a lower priority value will be used before features with a higher priority value. Only if all relevant Call Features of a lower priority are DOWN will a higher priority number Call Feature be used. It is recommended to use bigger numbers when setting up priorities (i.e. 10, 20, 30 instead of 1, 2, 3) to allow for inserting Call Features inbetween existing priority levels. For more information on how BVR uses Weight and Priority to provide load balancing and redundancy, please refer to Weight and Priority usage in the Blueworx Platform |
| Down Retry Interval | The amount of time before trying to use a Call Feature again once it has been determined to be DOWN. BVR will attempt to use it again sooner if no alternative options exist for the task on the application. |
| HTTP Proxy Group | The HTTP proxy group that will be used to make http/https requests on behalf of this call feature. If no proxy is currently up, we will reattempt using the highest priority one. If not set, no proxy will be used. |
| Timeout | The timeout BVR will use on messages to this call feature to abort the attempt and report an error. If set to 0s, BVR will wait indefinitely. Defaults to 0s. |
Overriding Parameters with CCXML
CCXML supports supplying hints on a <dialogstart> element to override most of the Eleven Labs specific parts of these settings for a particular VoiceXML session. See Overriding Eleven Labs TTS settings using hints
Please note that we do not support overriding the Xi API Key.
Note that Eleven Labs does not support SSML tags in the text it converts to audio, so any VoiceXML using an Eleven Labs TTS call feature must not use SSML tags in the <prompt>s.