Skip to main content

Chat APIs Overview

Vectara's Chat APIs provide a streamlined solution for integrating chatbot functionalities into domain-specific applications and websites using Retrieval Augmented Generation (RAG). Designed with efficiency in mind, these Chat APIs enable developers to create interactive user experiences and manage data conversations for users. With the unique conversationId generated for each chat session, developers can track and manage conversations over time.

These Chat APIs enable operations and pagination through the history of a chat, specifically the conversations and their subsequent exchanges, also known as turns. Administrators can gain valuable insights into user behavior by analyzing trends among users, which creates opportunities to identify knowledge gaps for information about products, technical documentation, and FAQs.

Vectara Chat provides the following Chat APIs:

  • Query contains a chat object within the summary which then has a unique conversationId.
  • List Conversations and get an an overview of the interactions between chatbots and users.
  • Read Conversations and retrieve detailed information about specific conversations by their IDs from a chat history corpus.
  • Delete Conversations by their specific Conversation IDs for data management and other purposes.
  • Delete Turns from a conversation starting from a specific Turn ID to manage the content of conversations.
  • Disable Turns to exclude turns from being used in generating responses.

Chat Object​

The summary within a Query contains the chat object which then specifies the conversationId and store status as true or false. Chats are set to false by default so you must set them to true or enable the chat option in the Vectara Console.

"chat": {
  "store": true,
  "conversationId": "string"
}

Conversation​

Conversations represent individual chat sessions, and a conversation starts with a chat request to the Query endpoint. A unique conversationId is generated at the initiation of the chat session, which serves as the identifier for all subsequent turns within this conversation. Vectara stores conversations in a single chat history corpus in the customer account.

{
  "conversation": [
    {
      "id": "ID of the conversation",
      "turn": [
        {
          "id": "ID of the turn",
          "conversation_id": "ID of the conversation",
          "query": "First query of the turn",
          "answer": "First answer of the turn",
          "enabled": true,
          "epoch_secs": 0
        },
      ]
    }
  ] 
}

Turns​

Each conversation has a series of turn objects, which are the sequence of message and response pairs that make up the dialog. Each turn represents a question/answer pair between the user and the chatbot and has a unique id that specifies its conversation_id.

"turn": [
  {
    "id": "ID of the turn",
    "conversation_id": "ID of the conversation",
    "query": "First query of the turn",
    "answer": "First answer of the turn",
    "enabled": true,
    "epoch_secs": 0
  },
  {
    "id": "ID of the second turn",
    "conversation_id": "ID of the conversation",
    "query": "Second query of the turn",
    "answer": "Second answer of the turn",
    "enabled": true,
    "epoch_secs": 0
  },
  // Additional turn IDs are created for each query and answer pair in the conversation
]