I created a chatbot and now I want to host it in a place that lets external users access it. I cant use Google cloud services as I do not have a credit or debit card which is acceptable by google and thus I can’t set up my billing account. And also I don’t want to pay for any services. Is there any other easy way I can do it. I am a complete noob. Please Help.
You can try to use AWS Free Tier to create e.g. an EC2 instance on that you’ll host your chatbot.
As far as I know, AWS (basically any other cloud/hosting provider) also requires to attach a debit/credit card to an account even if you use free tier.
If I want to host it only in a Intranet…is there anyway?
Hi @Swarnav, I hosted my bot on heroku for free. I used docker for this. It is a bit time consuming as you have to understand how both docker and heroku works. And it is also not very efficient being on heroku free tier hosting. But since it was a hobby project and dosn’t require you to have to enter a credit/debit card details, you might want to try that option.
Hi @_sanjay_r, could you please teach me or send me a link of how to deploy the bot to heroku? I have already tried, but all I get are errors. I’ve been using google server, but I’d like to learn how to properly deploy to heroku too. Thanks for your help!
I try with this one
but I have a problem with the slug size of my chat. and I couldn’t deploy it there
I hope it work for you
Continuing the discussion from Hosting My Chatbot:
Hi @marcos.allysson ,
Apologies for the very late reply.
I did this almost a year back and there were no direct links on how to achieve this, I somehow figured it out. Not sure about now or if there is a better way.
The way I deployed rasa bot to heroku, is with help of docker and github actions. Just understanding the basics is quite enough. Once your bot is ready, you need to have DockerFiles for both the NLU part and action server. It’s pretty staright forward till this and you can find Tutorials for this on YT, blogs etc.
Then specific to heroku, you need to add a heroku.yml file in the git repo(This is the file that heroku reads first and runs as configured)
Finally, add a workflow.yml file and configure it to deploy the bot on push/pull to the repo.(This is the github actions part. This allows us to overcome the slugsize exceeded issue).
Please refer to this repo to see the implementation, a bot which I hosted in a similar fashion. https://github.com/sanjayrjs16/CoronaSafeBot-Heroku