Offline installation in production server

Hi i am new to Rasa and i want to install for chatbot project to be hosted in our offline production server (not connected to internet). How can i completely install and setup in my offline production server.

Thanks in advance.

Hello @ak19,

First off welcome to the community!!

I would recommend testing out the Rasa X deployment as it comes with a lot of the stuff you need already out of the box like NGINX, etc.

You can find more information on setting it up from scratch at Installation and Setup and there is also a nice starter doc at Getting Started with Rasa

Let me know if you have any issues testing that out.

thanks for your response. I followed the documents of rasa installation and setup. Since Rasa requires additional dependencies. I want to install and setup rasa environment in offline windows server. How can i download whole package with dependencies and install it to my off the internet server.

One way could be to install rasa and train model on a machine with internet connection using docker and then take a docker dump of the containers with docker save
Transfer the containers on the offline server using a thumb drive and start containers using zip files by docker load.

You could also just download the docker images locally and run it that way.

Thanks for your replies @vp109 and @btotharye. The problem in docker is as per docker documentation it can run in windows 2016 and above. But my production server is having windows 2012 R2 installed. How can run docker images there. Is there any other option for creating the bundle for rasa installation files and its all dependencies at once.

Probably the only thing I can think of is maybe running a vm in that server but that is pretty hacky. The main architecture of Rasa X is setup via docker.

You could install all the relevant dependencies on a virtual environment and copy the files over to another server and activate the environment.
Although one issue with this would be since most of the python packages are platform dependent, you will have to make sure the place you’re creating the virtual environment should be running the same OS with same python version and take care of 32 or 64 bit architectures as well.

Another way to do it would be to download all the .whl files for dependencies and install them using pip but in that case you’ll have to download all relevant wheels one by one or you’ll have to look for a wheel with contrib