Newbie question: when I set up Verba to run in Replit it installs and runs OK, with a message “Port 8000 opened on host 127.0.01. Use the host 0.0.0.0 to open the port to web traffic” and the Replit Web Viewer says “Start a server to see the hosted output.”. I guess the web viewer is pointing at the wrong port. Is there a port setting I can change in a Verba file to fix this? Thanks for any suggestions.
Welcomet to our community
I have never tried Replit myself, but this is a really nice idea.
We will definitely look into this and bring share our findings here.
From what I tested, the
verba run command is not passing this arguments property in Replit as it does locally, so the host ends up being 127.0.0.1 no mater what you change.
Thanks for this idea! Very!
Thanks Duda. Colab or similar would be equally good. Any idea how long? Days/Weeks – it’s just that I have a ton of these things to test and I need to schedule. Thnx.
Let us know if you got it working @Mike_Hudson – would love to see Verba run on Repl.it
Hi @Mike_Hudson !
Sorry for the delay here. I had to play with Replit some more
I was able to make it running. Here is how:
You will edit your .replit file. Comment the
entrypoint directive, and set those two directives:
compile = ["pip", "install", "goldenverba"]
run = [
This should be enought to install and make it run.
I should craft a Verba template in the following days. I have not deployed it yet.
Please, let me know if this helps
Thanks Duda. This now runs OK in the Replit dev environment but I was unable to deploy. Were you able to do so?
I didn’t advanced to deployment.
Our team is planning to come up with a template to ease this process in Replit.
One thing I noticed was because Verba will also spin a embedded Weaviate server that binds to some ports, sometimes replit used those ports are the ones to be exposed.
So maybe if you point it to an external Weviate server (or WCS), it should always work.
As soon as we get news on that template I will update it here.