A comprehensive collection of recipes for building conversational AI assistants with Rasa's Conversational AI with Language Models (CALM) approach.
Rasa CALM Cookbook Structure:
.
├── .env.example
├── .gitignore
├── Makefile
├── pyproject.toml
├── README.md
└── recipes
├── level-1-basic
│ └── basic-tutorial
│ ├── .env
│ ├── .env.example
│ ├── actions
│ │ └── actions.py
│ ├── config-azure.yml
│ ├── config-local.yml
│ ├── config-openai.yml
│ ├── conversations
│ │ └── sample_conversations.md
│ ├── data
│ │ └── flows.yml
│ ├── domain.yml
│ ├── endpoints.yml
│ ├── Makefile
│ ├── pyproject.toml
│ ├── README.md
│ └── tests
│ └── e2e_test_cases.yml
├── level-2-intermediate
│ ├── enterprise-search
│ │ └── README.md
│ ├── multi-llm-routing
│ │ └── README.md
│ └── voice-assistant
│ └── README.md
└── level-3-advanced
├── coexistence-migration
│ └── README.md
├── deployment
│ └── README.md
└── fine-tuning
└── README.md
16 directories, 25 files
-
Clone the repository
git clone https://github.com/rodriveraai/rasa-calm-cookbook.git cd rasa-calm-cookbook -
Set up development environment
make setup source .venv/bin/activate -
Choose a recipe and get started
cd recipes/level-1-basic/basic-tutorial make setup-recipe
- basic-tutorial: Money transfer assistant following the official Rasa tutorial
- custom-actions: Building assistants with external API integrations
- voice-assistant: Voice-enabled assistant with speech recognition and synthesis
- enterprise-search: RAG-powered assistant with document retrieval
- multi-llm-routing: Cost-optimized assistant with multiple LLM providers
- fine-tuning: Fine-tuning smaller LLMs for improved performance and cost reduction
- deployment: Production deployment with Docker and Kubernetes
- coexistence-migration: Migrating from NLU-based assistants to CALM
- Python 3.10 or 3.11
- uv package manager
- Valid Rasa Pro license
- OpenAI API key (or alternative LLM provider)
Each recipe is self-contained with its own dependencies and can be run independently. See individual recipe READMEs for specific setup instructions.
- Fork the repository
- Create a new recipe following the established structure
- Test your recipe thoroughly
- Submit a pull request
This project is licensed under the AGPL-3.0 License - see the LICENSE file for details.