Conversations

Introduction

In Index Network, you can engage in conversations with one or multiple Indexes. This interaction is facilitated by our storage system, which processes and stores information to allow natural, meaningful dialogue. Whether it's querying for specific information or engaging in an ongoing conversation, the Index Network enables seamless communication with its indexed data.

Chat Streaming with SDK Client

To initiate a chat stream using the Index Network SDK Client, use the chat method. This method allows you to interact with the Indexes conversationally.

JavaScript

const queryResponse = await indexClient.chat({
  id: uuidv4(), // Provide a unique chat id for the query
  messages: [
    {
      content: "How do you evaluate a startup?",
      role: "user",
    },
  ],
  indexes: [indexId], // You can add all the indexes of a user as well
  // did: '...' // Optionally, provide a did
});


// Response is:
// {
//     answer: "This approach can be applied across various ..."
// }

Python

import uuid

chat_id = str(uuid.uuid4())

messages = [
    {
        "content": "How do you evaluate a startup?",
        "role": "user",
    },
]

response = client.chat(chat_id, messages, index_id)

print(response)

# Response is:
# {
#    answer: "This approach can be applied across various ..."
# }

API Endpoint Reference

The corresponding API endpoint for initiating a chat with Indexes is described below:

Post chat data

POSThttps://dev.index.network/api/discovery/chat

Body

application/json

id*string

Unique identifier for the chat.

messages*array of string

Array of messages.

didnullable string

Optional decentralized identifier.

typenullable enum

Type of the chat, valid if 'did' is provided.

ownedstarred

indexIdsnullable array of string

Optional array of index identifiers.

Response

200400

Successful response with streamed answer

Body

application/json

statusstring

Status of the response.

dataobject

Request

JavaScriptCurl

const response = await fetch('https://dev.index.network/api/discovery/chat', {
    method: 'POST',
    headers: {
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      "id": "text",
      "messages": [
        "text"
      ]
    }),
});
const data = await response.json();

Response

{
  "status": "success",
  "data": {
    "answer": "This is the streamed answer to your question."
  }
}