Rasa

Rasa is an open-source framework for building conversational assistants (chatbots, voicebots) offering advanced natural language understanding (NLU), stateful dialogue management, and the ability to fully self-host the bot. The core is free with no deployment limits: a single organization can use one copy of the code in production internally or on cloud. The platform covers two flows: intent recognition, entity extraction, and custom dialogue flows via YAML files or a flow builder interface (Rasa Studio). The free edition is limited to 1,000 conversations/month in external interactions or 100 internal conversations.

Rasa offers no French-language interface, though the framework supports multi-language for NLU. A complete API enables integration into any backend. NLU models can be trained on your own data to keep conversations internal (GDPR/data residency by default if self-hosted). Rasa targets primarily tech teams; the hidden cost is time-to-value: implementation, model tuning, and infrastructure management require a dev team. Paid offerings Business (custom pricing) cover no-code Studio, Rasa Pro (managed infrastructure), and premium support.