Rasa is an open-source machine learning framework for building text- and voice-based conversational systems. It provides natural language understanding (NLU) for intent classification and entity extraction, along with dialogue management to handle multi-turn conversations. Developers use it to create chatbots and voice assistants that connect to platforms like Slack and Facebook. The project is currently in maintenance mode, with future development shifting to Hello Rasa and the CALM (Conversational AI with Language Models) engine.