simple_rag_javascript

The Developer's Guide to AI: From Prompts to Agents

Create a simple RAG solution using JavaScript

Most LLM-powered applications use RAG in some way.

The key to RAG is getting your data to the LLM. This tutorial shows you how to use a vector search to do just that.

Remember, we're limited on the amount of data we can send to the LLM:

• prompts have a token limit
• LLM providers often charge by the token
• including irrelevant data may confuse the LLM response

We need an efficient way to find only relevant data.

Vector search is a great choice. We prompt LLMs with natural language and a vector search allows you to find relevant data based on the meaning of what's being asked.

Quick Start

The following steps will get you up and running on your machine.

0. Install Ollama and run:

• Download and install: https://ollama.com
• Verify the install and start Llama in a terminal window:

~ % ollama run llama3.2

1. Clone this project into a local directory:

~ % git clone <url>

2. Install Node

• Download and install: https://nodejs.org
• Verify the install in your terminal:

~ % node -v

• If the installation succeeded, the version will print

3. Install node modules

• Navigate to the project part3/simple_rag_javascript directory in your terminal and run:

~/developers-guide-to-ai/part3/simple_rag_javascript % npm install

4. Launch the server

• Back in the terminal, run the following commands:

~/developers-guide-to-ai/part3/simple_rag_javascript % node server.mjs

5. Launch the client

• In a terminal, navigate to the part1/client directory and run the following commands:

~/developers-guide-to-ai/client % npm run dev

6. Open your web browser and visit: http://localhost:5173

• Type the question into the input: Can you tell me how to login to Workday?
• The article will be retrieved and passed to the LLM to generate a response