Warning: This document is for the development version of Rasa X. The latest version is 0.24.1.

Improve your contextual assistant with Rasa X

To build great conversational AI, you need to improve your assistant based on real conversations. You can’t anticipate how users will talk to your assistant, so there is no substitute for learning from their conversations as soon as possible.

Rasa X is a toolset that helps you leverage conversations to improve your assistant. Those conversations can be with you, testers, or real users. As soon as your assistant can handle the most important happy path stories, Rasa X helps take it to the next level.

We created Rasa X because some things — like reviewing conversations and turning them into training data — are much easier with a user interface. Rasa X focuses on those workflows, enabling the things that are hard to do in the command line.

Rasa X

  • layers on top of Rasa Open Source and helps you build a better assistant
  • is a free, closed source toolset available to all developers
  • makes it easier to review your conversations and figure out how to improve your assistant
  • is not tied to a specific third-party cloud service
  • can be deployed anywhere, so your training data stays secure and proprietary

So, what can you do with Rasa X?

Enable workflows: easily capture conversations with your assistant

Talk to your assistant: chat via Talk to your bot or use Interactive Learning to teach it as you chat.

Conversations with test users: Share with guest testers to allow others to quickly try out your assistant.

Conversations with real users: connect to a channel and view real conversations with users.

Review conversations: filter conversations as you view them to figure out what to do next

Once you start collecting a lot of conversations, it can be difficult to process every conversation. With Rasa X, you can filter conversations to focus on those most likely to need your attention.

As your team reads through the conversations and discovers the weak points of your assistant, you can also flag messages and share conversations to more easily collaborate with each other.

Improve assistant: continually update your assistant in Rasa X and on the command line

As part of the Rasa X server setup, you connect your server to a remote Git repository to load your assistant into Rasa X. You can continue to use this connection to ensure any changes you make are reflected in Git. This allows you to version control your assistant and makes the process of making updates both on the Rasa X server and on your local machine smooth.

This is key because in order to improve your assistant, you will need to make incremental updates inside and outside of Rasa X based on what you learn from your conversations.

As part of your improvement workflows, you can also use your favorite Git-based automation server to set up a CI/CD pipeline to run end-to-end tests, model evaluations, and integration tests. This will ensure you do not introduce model regressions and will give you the confidence to deploy a new model.

Install & Setup

Deploy Rasa X to a server and load your assistant in using Integrated Version Control to start improving it!