Configuration Parameters

20min
Common parameters are attributes of the Digital Human that are available across all  Operating Modes﻿. These are used to further personalise your Digital Human to achieve your desired end user experience. 
One or more parameters can be changed within a single Digital Human update as specified in the Update a Digital Human﻿ documentation.
Below you can find the parameters listed: 
The parameters listed here are commonly updated atleast once after the initial Digital Human creation process. 
Conversation
greetings
This is the welcome message that the Digital Human will deliver to the end user upon first interaction. 
Type: string
Example "Hi there, I am john"
Operating Modes: applies to all
suggestions 
These are a list of menu-like options that suggest to the user what to engage the digital human about. These options are clickable via the UNITH front end. 
Type: list of strings. 
Example ["Who is UNITH", "Where are UNITH offices?"]
Operating Modes: doc_qa, oc
In doc_qa mode, the list of suggestions is dynamically generated when you upload a document. It is highly recommended to review and edit the list to align with your use case. Furthermore, every time you make issue a Document Upload, the suggestions list is over-ridded with a new dynamically generated list. Therefore, it is advised to keep a local copy of your suggestions list, before you upload the document.
Suggestions can only be actioned once in a given conversational session.
isRandomSuggestions
Informs UNITH if the suggestions should be delivered at random the end user or if they will be sequenced in order of the list 
Type: true/false
Operating Modes: doc_qa, oc
Personalisation
alias
The name of the Digital Human as it appears to the end user engaging with the Digital Human. 
Type: string
Example "David Beckham"
Operating Modes: applies to all
name
The name of the Digital Human used only for classification purposes inside the UNITH platform. Not recommended to change. 
Type: string
Example "beckham"
Operating Modes: applies to all
TTS & STT
languageSpeechRecognition
Language code to improve speech recognition. Choose a value supported by Speech service﻿
Type: string
Example "en-US"
Operating Modes: applies to all
ttsProvider
Name of the Text-To-Speech audio synthesis engine UNITH platform will leverage. Only "audiostack" is currently supported. 
Type: string
Example "audiostack"
Operating Modes: applies to all
ttsVoice
The voice that will be associated with the Digital Human. This list is sourced from AudioStack. You can access the list directly via UNITH API as detailed List Voices﻿﻿
Type: string
Example "kelso"
Operating Modes: applies to all
customWords 
A dictionary containing pairs of words and their custom pronunciation. You can use this to override how a word is usually pronounced. Can be empty.
Type: dict
Example "customWords": {"UNITH":"Unith", "Idan":"eedan", "Xavi":"Chavi"}
Operating Modes: applies to all
customWords are specific to the ttsVoice. A custom word that is required for one voice, may not be required for another voice. Hence it is important to consider your customVoice for each Digital Human you create, independently. 
phrases
A list of phrases used to provide better speech recognition. This is often brand names and uncommon proper nouns that can be difficult to recognise without prior knowledge. Can be empty.
Type: list of strings. 
Example "phrases": ["Unith", "Casio", "Idan"]
Operating Modes: applies to all
UI
language
UNITH front end application language. Current supported languages are ar-AE, bs-BA, cs-CZ, en-US, es-ES, fr-FR, lt-LT, lv-LV, nl-NL, pl-PL, pt-PT, ru-RU, sk-SK, sr-RS
Type: string
Example "language": "en-US"
Operating Modes: applies to all
When CREATING the Digital Human, "langCode" is used, instead of "language"
﻿
Update Digital Human Example
Per description in Update a Digital Human﻿, the below is an example payload that is executed against the PUT head/update endpoint.
JSON
{
  "id": "asdfasf-1233-1123-b3fc-4d563d56a3a6",
  "name": "john",
  "alias": "John",
  "ttsVoice": "kelso",
  "greetings": "Hi there, I am Jonny",
  "isRandomSuggestions": false,
  "phrases": [
    "Unith", "Casio", "Idan"
  ],
  "customWords": {"UNITH":"Unith", "Idan":"eedan", "Xavi":"Chavi"}
{
  "id": "asdfasf-1233-1123-b3fc-4d563d56a3a6",
  "name": "john",
  "alias": "John",
  "ttsVoice": "kelso",
  "greetings": "Hi there, I am Jonny",
  "isRandomSuggestions": false,
  "phrases": [
    "Unith", "Casio", "Idan"
  ],
  "customWords": {"UNITH":"Unith", "Idan":"eedan", "Xavi":"Chavi"}
﻿
Operational Mode Parameters
operationMode
The converational mode that the Digital Human will employ. 
Type: fixed value string
doc_qa: Ideal for document-informed Questions & Answers. Leverages UNITH RAG pipeline. Used to power Restricted Open Conversations﻿﻿
oc: Open Conversation, purely based on ChatGPT
ttt: Used for Text-to-Video
plugin: Used to leverage another conversational engine instead of UNITH
voiceFlow: Used for guided conversations built with voiceflow
See Operation Modes definitions﻿﻿
ttt
oc
doc_qa
voiceflow
plugin
{
    ...
    "operationMode": "ttt",
    ...
} 
{
    ...
    "operationMode": "ttt",
    ...
} 
﻿
UNITH does not support the updating of operationModes. If the operationMode needs to be changed, you will need to create a new Digital Human. 
For voiceflow, the "greetings" welcome message must be defined in Voiceflow.
﻿
﻿
﻿
Updated 02 Dec 2024
Did this page help you?
Update a Digital Human
Overview