- Why Rasa?The standard infrastructure layer for conversational AI - open source
- FeaturesBuild, improve, and deploy contextual chatbots and assistants
- Pricing
- Support Options
- Join the CommunityBe a part of the fastest growing conversational AI community
- How to ContributeGet involved! There are lots of different ways to contribute
- View all Customer Stories →
Docs

Create assistants that go beyond basic FAQs

Machine learning tools for developers to build, improve, and deploy contextual chatbots and assistants. Powered by open source.

Developers at leading companies use Rasa

Add context to your chatbot

Automate text-and voice-based conversations end-to-end, integrate with your APIs, and connect to any channel.

Join our fast-growing developer community

Rasa Community is a diverse group of makers and conversational AI enthusiasts.

500k+

Downloads

3,500+

Forum Members

300+

Contributors

Customer Stories

Automating 30% of customer service requests

N26, the leading mobile bank in Europe uses their in-app AI assistant Neon to offer a faster service to their users. Learn more about how they use Rasa to automate their customer support delivering the best experience.

As seen on:

From the blog

View all posts

Integrate response retrieval models in assistants built with Rasa

by Daksh Varshneya on Sep 12, 2019

Do you want to handle a lot of FAQ, chitchat and other single-turn interactions within your assistant built with Rasa? With Rasa 1.3.0 released, it’s easier than ever. Check out the blogpost to find out how.

Rasa Developer Summit - Overview

by Alan Nichol on Sep 10, 2019

We are thrilled to be hosting the second Rasa Developer Summit [https://rasa.com/summit/] in just 2 weeks in San Francisco. We have more than 20 talented speakers [https://rasa.com/summit/agenda/] from leading entities in conversational AI - companies, consultancies, academe, and technology providers. Each of the presentations is unique and filled with plenty to learn, but fall roughly into one of 5 categories: * Practical Rasa Implementations * Challenges in Creating Contextual Assistants

Pruning BERT to accelerate inference

by Sam Sucik on Sep 5, 2019

After previously discussing various ways of accelerating models like BERT [https://blog.rasa.com/compressing-bert-for-faster-prediction-2/], in this blog post we empirically evaluate the pruning approach. You can: * read about the implementation of weight and neuron pruning with BERT (jump to the section), * see how much faster and smaller we can make BERT (jump to the section), * understand what neuron pruning can tell us – e.g that sometimes BERT doesn’t need attention at all (jum