# Rasa

> Enterprise conversational AI platform for building reliable AI agents with LLMs, business logic, and full control over behavior and performance.

Rasa is an enterprise-grade conversational AI platform that enables organizations to build reliable AI agents capable of handling millions of conversations. The platform extends large language models (LLMs) with structured business logic, deterministic flows, and built-in recovery patterns, giving teams complete control over agent behavior and performance. Trusted by leading enterprises like Autodesk, Swisscom, and T-Mobile, Rasa powers customer experience, support, sales enablement, and operational efficiency use cases across industries including finance, healthcare, retail, and telecom.

- **CALM Architecture** combines the flexibility of LLMs with deterministic logic, keeping business rules separate from language understanding for predictable, testable agent behavior.
- **Enterprise RAG** retrieves information in real time from your trusted data sources, ensuring every answer is fresh, verifiable, and aligned with your knowledge bases.
- **Voice Gateway** enables seamless voice experiences with built-in turn-taking, timeouts, and latency control for IVR and voice AI deployments.
- **Natural Language Understanding (NLU)** provides language-agnostic intent and entity recognition with full control over multilingual pipelines.
- **Agentic AI** creates agents that take initiative and adapt to complexity while maintaining guardrails and control.
- **Orchestration** coordinates multiple agents and tools to deliver the right help at the right moment across complex workflows.
- **MCP Integration** gives AI agents a standard way to connect with APIs as tools using the Model Context Protocol.
- **Rasa Studio** offers a no-code interface for business users to build, analyze, and optimize AI assistants with flow builders and content management.
- **Multi-LLM Management** supports fine-tuning recipes and smart LLM usage to optimize cost and performance.
- **Enterprise Security** includes PII data management, secrets management via Vault, SSO, RBAC, and dependency vulnerability protection.

To get started, developers can request a free Developer Edition license for local or production use, or try the Hello Rasa demo. The platform supports deployment via Kubernetes with Helm charts, offers REST and WebSocket channel connectors, and integrates with observability tools through OpenTelemetry. Enterprise teams benefit from premium support with dedicated Customer Success Managers and Engineers.

## Features
- CALM (Conversational AI with Language Models) dialogue management
- Enterprise RAG with real-time retrieval
- Voice Gateway with IVR support
- Language-agnostic NLU
- Agentic AI with guardrails
- Multi-agent orchestration
- MCP integration for API tools
- No-code flow builder (Rasa Studio)
- Multi-LLM management and fine-tuning
- PII data management
- SSO and RBAC
- OpenTelemetry observability
- Kubernetes deployment with Helm
- End-to-end testing
- Custom actions server (Rasa SDK)
- Channel connectors (REST and WebSocket)
- Multilingual AI support
- Contextual response rephrasing
- Enterprise search
- Secrets management with Vault

## Integrations
Kubernetes, Helm, Redis, Kafka, Vault, OpenTelemetry, AudioCodes VoiceAI Connect, REST API, WebSocket

## Platforms
LINUX, IOS, WEB, API, DEVELOPER_SDK

## Pricing
Freemium — Free tier available with paid upgrades

## Links
- Website: https://rasa.com
- Documentation: https://www.rasa.com/docs
- Repository: https://github.com/RasaHQ
- EveryDev.ai: https://www.everydev.ai/tools/rasa