What is this?
Rasa Open Source is an open source machine learning framework for automating text- and voice-based conversations — you build contextual assistants like chatbots and voice assistants, then connect them to Slack, Messenger, Telegram and other channels.
Rasa Open Source is an open source machine learning framework for automating text- and voice-based conversations — you build contextual assistants like chatbots and voice assistants, then connect them to Slack, Messenger, Telegram and other channels.Is it a framework I code with, or more of a hosted chatbot builder?It's a framework: a Python codebase you use to build your own assistant, from NLU to dialogue management. The enterprise Rasa Platform is a separate product — this repo is the open source framework itself.
It's a framework: a Python codebase you use to build your own assistant, from NLU to dialogue management. The enterprise Rasa Platform is a separate product — this repo is the open source framework itself.What do I need installed to run or hack on the classic framework?A Python environment — the project is Python end to end and uses Poetry for packaging and dependency management. A Dockerfile ships in the repo, and macOS users building from source may need to set SYSTEMVERSIONCOMPAT=1 if the compiler complains.
A Python environment — the project is Python end to end and uses Poetry for packaging and dependency management. A Dockerfile ships in the repo, and macOS users building from source may need to set SYSTEMVERSIONCOMPAT=1 if the compiler complains.Is this repo still the way to build with Rasa today?Not quite — Rasa Open Source is in maintenance mode. The project's focus has moved to Hello Rasa and the CALM engine, so expect fixes here rather than new features.
Not quite — Rasa Open Source is in maintenance mode. The project's focus has moved to Hello Rasa and the CALM engine, so expect fixes here rather than new features.Where can an assistant built with it actually live?It connects to Facebook Messenger, Slack, Google Hangouts, Webex Teams, Microsoft Bot Framework, Rocket.Chat, Mattermost, Telegram and Twilio, plus your own custom conversational channels.
It connects to Facebook Messenger, Slack, Google Hangouts, Webex Teams, Microsoft Bot Framework, Rocket.Chat, Mattermost, Telegram and Twilio, plus your own custom conversational channels.What is CALM?CALM (Conversational AI with Language Models) is Rasa's newer engine: the LLM understands what the user says while code enforces the business rules. You describe flows as logical steps — like collect money or transfer funds — instead of rigid dialogue trees.
CALM (Conversational AI with Language Models) is Rasa's newer engine: the LLM understands what the user says while code enforces the business rules. You describe flows as logical steps — like collect money or transfer funds — instead of rigid dialogue trees.Does the Hello Rasa flow still need intent training data?No. The README says they have moved beyond intents: the LLM handles dialogue understanding while you define the business flows, and a built-in copilot helps generate code, debug flows, and expand your agent.
No. The README says they have moved beyond intents: the LLM handles dialogue understanding while you define the business flows, and a built-in copilot helps generate code, debug flows, and expand your agent.Is Hello Rasa production-ready or just a demo?The README calls it production ready — 'not just a toy' — and you can export your agent to the Rasa Platform when you are ready to scale. An Inspector shows real-time decision-making, no black boxes.
The README calls it production ready — 'not just a toy' — and you can export your agent to the Rasa Platform when you are ready to scale. An Inspector shows real-time decision-making, no black boxes.Is there a community around this?Yes — a vendor-neutral Agent Engineering Community for people shipping real-world AI agents, with channels like #agent-design for architecture discussions, #showcase for demos, and #ask-anything for debugging.
Yes — a vendor-neutral Agent Engineering Community for people shipping real-world AI agents, with channels like #agent-design for architecture discussions, #showcase for demos, and #ask-anything for debugging.Which topics is it in?bots, chatbots, conversational-ai, voice, bot, ai-assistant, slack, natural-language-processing, telegram, nlp, chatbot, framework
bots, chatbots, conversational-ai, voice, bot, ai-assistant, slack, natural-language-processing, telegram, nlp, chatbot, framework