ChatLab vs Rasa side by side.

Rasa is the serious open-source choice. Apache 2.0 licensed, fully self-hostable on your own infrastructure, a customizable NLU pipeline, the CALM dialogue framework, and no vendor lock-in of any kind. For regulated environments where data cannot leave the building, that combination is hard to argue with.

It is a framework, not a product, and the comparison has to be read that way. Rasa does not train from website URLs, a sitemap or markdown files. There is no chat widget, no iFrame, no shareable link, no live chat, no lead collection, no conversation rating and no welcome screen. There is no avatar, chat icon, colour or launcher customization, no knowledge base optimizer, no domain restrictions, no automatic retraining, no AI conversation insights, no client or conversation summaries, no Offer Cards, no dynamic suggested follow-ups and no spam filter. All of it is yours to build, host and maintain.

Rasa itself is free. ChatLab is Free or $15/mo at $3.13 per 100 conversations, hosted and maintained. Rasa costs engineering salary instead of a subscription. If you have that team and a data residency requirement, it is the right answer. If you do not, it is by far the most expensive option here.