Hi Rasa @community!
We released Rasa X 0.26.0 last Friday
There are a lot of updates, so I thought I would summarize them for you. Check out what’s new…
1) Connect Rasa X to Git-based development workflows
Integrated Version Control
is now a core feature of Rasa X with a simplified setup. You can load your assistant into Rasa X via the UI by adding the SSH url of your remote Git repository, specifying which branch you want to use, and copying the public SSH key to the deploy key section of GitLab, GitHub, BitBucket, etc.
Docs: Connect Rasa Open Source to Rasa Enterprise
Blog: Connect in Fewer Steps: What’s New with Integrated Version Control | The Rasa Blog | Rasa
2) Focus on the conversations most likely to need your attention with new filters
All conversations you have with your assistant in Rasa X now show up in Conversations. But don’t worry, they won’t clutter the screen because you can now filter by channel
. There are also additional new filters
for slots, confidence, and stories not in your training data.
Docs: Review Conversations
Blog: Reviewing Conversations in Rasa X 0.26 | Rasa Blog | The Rasa Blog | Rasa
3) Tag as you review conversations to pattern match across many conversations
When you are reviewing conversations, you can now use tags via the API
to help measure success and tags in the UI
to help you track conversation patterns that require human judgement to identify, like instances where users struggled or requested a specific feature.
Docs: Review Conversations
Blog: Using Conversation Tags to Measure Carbon bot’s Success Rate | The Rasa Blog | Rasa
4) Deploy your assistant in minutes with a setup ready to scale
The easiest way to deploy an assistant is to now use our One-Line Deploy Script that sets you up with a scalable Kubernetes cluster. This method deploys both Rasa X and a production-ready Rasa Open Source container, so you’re ready to go live by loading in your assistant with Integrated Version Control.
Docs: Installation Guide
Blog: 1. Create a VM & 2. curl -s http://get-rasa-x.rasa.com | sudo bash
5) Forward messages from your existing Rasa assistant to Rasa X
If you are not using Rasa X with your assistant yet, we have made it really easy to connect it to your existing deployment
. There is even a rasa export
command that lets you send all of your past conversations to Rasa X.
Docs: Connect Rasa Open Source to Rasa Enterprise
Blog: How to Migrate Your Assistant to Rasa X (the Easy Way) | The Rasa Blog | Rasa
Wait, what’s Rasa X? When should you use it?
Rasa X is a toolset that helps you leverage conversations to improve your assistant. It layers on top of Rasa Open Source and helps you build a better assistant by making it easy to to turn real conversations into training data.
You use Rasa X once you have a minimum viable assistant, one that can handle the most important happy paths. Following “the loop” allows you to build the best possible assistant.
Docs: Introduction to Rasa X/Enterprise
Blog: Rasa Open Source + Rasa X: Better Together | Rasa Blog | The Rasa Blog | Rasa
If you have thoughts about how we could improve Rasa X or would be willing to chat about your experience good or bad, let me know at t.dunn@rasa.com!