Rasa, founded in 2016 by Alan Nichol and Alex Weidauer, builds an open developer platform for enterprise AI agents rather than a closed, fully-managed SaaS product, giving a team full access to prompts, policies, and the underlying codebase instead of a black box that produces answers without visibility into why.
Its CALM dialogue-management system extends LLM flexibility with deterministic business logic a team can inspect and control directly, and the platform supports on-premises, private-cloud, and fully air-gapped offline deployment, a deployment option most cloud-only competitors in this category don't offer at all. A team also chooses its own LLM and infrastructure rather than being routed through one provider by default.
A free Developer Edition covers standard use within conversation limits, and Growth-tier pricing starts around $35,000 a year for higher volume. For a technical team that wants full visibility into agent reasoning and the option to deploy fully offline or air-gapped, rather than a closed cloud-only platform, Rasa's open, inspectable architecture addresses that directly.
