Getting Started with Colang 2.0

This section is a getting started guide for Colang 2.0. It starts with a basic “Hello World” example and then goes into dialog rails, input rails, multimodal rails, and other Colang 2.0 concepts like interaction loops and LLM flows. This guide does not assume any experience with Colang 1.0, and all the concepts are explained from scratch.

Prerequisites

This getting started guide will focus only on the Colang files. For complete details on how to install NeMo Guardrails and create a sample configuration, refer to the Installation Guide and the NeMo Guardrails Getting Started Guide.

The config.yml file for all the examples should have the following content:

config.yml

1 colang_version: "2.x"
2 
3 models:
4   - type: main
5     engine: openai
6     model: gpt-4-turbo

The above config sets the Colang version to “2.x” (this is needed since “1.0” is currently the default) and the LLM engine to OpenAI’s gpt-4-turbo. Make sure to set the required API access key as an environment variable (e.g. OPENAI_API_KEY for OpenAI API). See section Supported Models for all supported models.

Check the section Development and Debugging for how you can install Colang syntax highlighting to make editing Colang scripts easier.

Terminology

At a high level, Colang adopts as much as possible from the Python terminology. This guide will talk about Colang scripts and modules (i.e., .co files), packages (i.e., folders), standard library, importing mechanism, etc.

Guides

Hello World

Create the first Colang 2.0 flow.

Dialog Rails

Add dialog rails with LLM integration.

Multimodal Rails

Use multimodal events and actions.

Input Rails

Add input rails to inspect and block user input.

Interaction Loop

Build a simple interaction loop.

LLM Flows

Let the LLM generate flow continuations.

Recommended Next Steps

Choose what to learn after the getting started guide.