In this tutorial you will create a new conversation with your contact center, publish text messages, and receive agent responses.


  • Docker
  • LivePerson account enabled with two features: Async_Messaging and Authenticated_Chat. If you are not sure that your account is enabled with these two features, please contact LivePerson Support (use the chat on this page, or message Support).

Step 1 — Launch your Shell

We will use a docker image called lp-shell to run the shell for this tutorial. It is a Linux image equipped with curl,wscat,jq and several scripts.

After replacing the __YOUR_ACCOUNT_ID__ with your account ID, type the following into your terminal:

docker run --env LP_ACCOUNT=__YOUR_ACCOUNT_ID__ -it lpinc/shell

You should receive a shell line.

Step 2 — Create a Token

Note: The below is only a demonstration of how to create an unauthenticated token. Please do not use the below method in your production environments. For production environments, only use the authenticated method as described here.

In this demonstration we will create a new unauthenticated consumer identity. Run the following command:

LP_JWT=`curl -X POST https://$LP_IDP/api/account/$LP_ACCOUNT/signup | jq -r .jwt`

This will request the creation of a new identity, extract the JWT from the response, and set it as a variable called LP_JWT in your shell. This new identity will be used in the next step.

To check the value inside the LP_JWT, type:

> echo $LP_JWT

Step 3 — Connect to the Messaging Service

Using the JWT we have created, we can open the connection to the UMS:

wscat -k 60 -H "Authorization:jwt $LP_JWT" -c "wss://$LP_ASYNCMESSAGINGENT/ws_api/account/$LP_ACCOUNT/messaging/consumer?v=3"

In the following steps we will send JSON requests and receive responses and notifications through this connection. If the connection is closed, you can reconnect using the same wscat command.

Step 4 — Create a New Conversation

Before creating the conversation, log into Conversational Cloud as an agent. In order for the agent to receive the messages, ensure that no other agent is logged into your account.

To request a new conversation, paste the following message into the connection from the previous step.


In response, you will receive the ID of the new conversation.

	"kind": "resp",
	"reqId": "1",
	"code": 200,
	"body": {
		"conversationId": "7507be78-60ef-4468-b3b1-baa47fbeea21"
	"type": "cm.RequestConversationResponse"

Open the Agent Workspace (from Step 1), and accept the incoming request by clicking the Accept button.


Step 5 — Publish a Text Message

In order to publish content to a conversation, substitute the __YOUR_CONVERSATION_ID__ with the conversationId you got in the response of Step 4, and paste it into the opened WebSocket.

{"kind":"req","id":2,"type":"ms.PublishEvent","body":{"dialogId":"__YOUR_CONVERSATION_ID__","event":{"type":"ContentEvent","contentType":"text/plain","message":"hi there"}}}

Note: You can use the message builder () to build the above message.

The published message will be displayed on the agent side: consumer-msg

Now close the conversation from the Agent Workspace using the Close conversation menu item: close

Step 6 — Record LP_JWT to Next Tutorials

In order to use the consumer token (LP_JWT) in the following tutorials, we will put a file on your machine. Create a new file outside of the lp-shell and paste the following content into it:

LP_ACCOUNT=qa57221676	put your account id here

* Put here the JWT you have created in this tutorial
* You can use echo $LP_JWT from your lp-shell to view it

What's next

You can now close the lp-shell and move on to the next tutorial.

Explore API messages in the API Reference page. In order to build your own messages, you can use the message builder