CognigyClient

Repo for the cognigy (server) client which can be used to connect to the Cognigy.AI platform from server applications written in Node.JS.

This client is only compatible with the 3.X version of our product and will no longer work with the old 2.X versions!

Installation

To install the cognigy client for your server project, use the following:

npm i @cognigy/cognigy-client --save

Usage

If you want to use the CognigyClient you have to create a Socket Endpoint on the COGNIGY.AI platform. For a detailed description of creating and configuring a Socket Endpoint see our Documentation.

Requirements

You'll need two bits of information from our COGNIGY.AI platform to sucessfully connect to your configured Socket Endpoint:

• The Base URL of the Socket Endpoint
• The URL Token of the Socket Endpoint

You can find them on the Endpoint Editor page of your Socket Endpoint.

Example

You get you started quickly, simply copy-paste this sample code and adjust your options where necessary.

import { CognigyClient, Options } from "@cognigy/cognigy-client";

const options : Options = {
    /** Required fields */
    baseUrl: 'server-address',
    URLToken: 'endpoint-url-token',
    userId: 'your-user-id',
    sessionId: 'unique-session-id',
    channel: 'channel-identifier',
    forceWebsockets: false,
    handleOutput: (output) => {
        console.log("Text: " + output.text + "   Data: " + output.data);
    },

    /** Optional fields */
    keepMarkup: true,
    reconnection: true,
    interval: 1000,
    expiresIn: 5000,
    passthroughIp: "127.0.0.1",
    handleError: (error: CognigyError) => { console.log(error); },
    handleException: (error: CognigyError) => { console.log(error); },
    handlePing: (finalPing: IFinalPing) => { console.log("On final ping"); }
};

const client: CognigyClient = new CognigyClient(options);
client.connect()
    .then(() => {
        client.sendMessage("I like pizza", { key: "value" });
    })
    .catch((error) => {
        console.error(error);
    });
``