Clai Free Edition

The Clai Free Edition is available for download to run within Docker containers. The docker-compose file and relevant supporting directories and configuration files are available via the pemari website at (COMING SOON).

To run the Clai Free Edition on a Windows PC as a test you will require a docker environment. You can use either Docker Desktop or an alternate method of running the docker containers on your PC, such as multipass or virtualbox and a linux VM with docker installed and configured.

The easiest option is to use Docker Desktop as Clai is configured to run within Docker Desktop on your local pc/laptop.

Clai Free Edition will download 4 images from docker hub:

nginx:1.21
mongo:5.0.4
pemari/clai-admin:1.2.8
pemari/clai-rasa:2.3.3.cl.13
When the containers are running the database files and trained models will be stored on the host file system so will not be lost if the container is downed.

Docker Desktop

To start up Clai proceed as follows:

Expand the clai.zip file to the clai folder
In a command prompt or shell change directory to the clai folder.
Run docker compose to download the necessary images from docker hub and start the containers:
```
  docker compose up -d
```
The first time starting up can take a little while as the required images will be downloaded to your PC from docker hub.
Once the containers are running, you can check in docker desktop then go in a browser open Clai admin in a browser. If this shows the welcome screen, the Getting Started Page will direct you on the next steps.

Docker running on a Linux Server/VM

Note: We test clai running within docker on an Ubuntu Focal (20.04) VM.

To run Clai you will need the docker engine and cli on your linux server/VM. The instructions for installing docker on a Linux host can be found at Install Docker Engine.

In addition you will need the docker-compose plugin for docker. The instructions for the download and installation of the compose plugin can be found on Github at Docker Compose and you will need to goto the release page to actually download the applicable executable.

Once you have docker ready then you can proceed with getting Clai up and running.

Expand the clai.zip file to the clai folder.
In the clai folder edit the file .env and change the variable ADMIN_ROOT_URL to reflect your linux server/vm, i.e.
```
ADMIN_ROOT_URL=http://[IP address or name]:8888
```
In the clai/nginx/html/webchat folder edit the file index.html and change the socketUrl (line 17) value from http://localhost:8888 to:
```
socketUrl: "http://[IP address or name]:8888/"
```
In a command prompt or shell change directory to the clai folder.
Run docker compose to download the necessary images from docker hub and start the containers:
```
docker compose up -d
```
The first time starting up can take a little while as the required images will be downloaded to your PC from docker hub
If your linux host is running a firewall then you will need to open the port for clai, i.e. 8888.
Once the containers are running, you can check in docker desktop then go in a browser open Clai admin in a browser at the ADMIN_ROOT_URL you defined in the .env file. If this shows the welcome screen, the Getting Started Page will direct you on the next steps.

Webchat

The webchat is available within Clai Admin but also as a standalone web page at http://localhost:8888/webchat or at the defined ADMIN_ROOT_URL/webchat. Initially the webchat will not respond with anything until some data is added in Clai Admin and the bot is trained. To connect the webchat to Clarity then see the section Clarity WebChat

You must now restart the rasa service or the rasa docker container, i.e. rasa-development. Now test by accessing:

Clarity at http://[CLARITYHOST:CLARITYPORT]/pmb
the standalone webchat at http://CLAIHOST:CLAIPORT/webchat
the Admin webchat within the project in the admin tool at http://CLAIHOST:CLAIPORT