Quick Start Guide

Set up a basic RAG-powered AI chatbot with ChatRAG in 12 simple steps

1

Access the Repository

After purchasing ChatRAG, you'll receive an invitation to a private repository (e.g., GitHub).

Clone the repository to your local machine:

git clone <repo-url>

Or download the codebase directly from the repository.

2

Install Dependencies

Open your IDE and navigate to the ChatRAG codebase directory.

npm install

This will install all required packages for your project.

3

Run the Development Server

In a new terminal window, start the development server:

npm run dev

The first run will generate a local configuration file where you will set all necessary environment variables.

4

Open the Configuration Dashboard

Run the configuration dashboard to easily manage API keys and features:

npm run config

This opens a visual configuration dashboard where you can input and manage API keys and toggle features.

5

Obtain and Paste API Keys

Acquire the following API keys and paste them into the configuration dashboard:

Llama Cloud

Register at cloud.llamaindex.ai and generate a secret key for document parsing (offered by LlamaIndex).

OpenAI

Generate an API key at platform.openai.com to be used for embeddings (and optionally for model access).

OpenRouter

Generate an API key at openrouter.ai to access multiple LLMs.

6

Set Up the Vector Database (Supabase)

Sign up or log in to Supabase and create a new project:

  1. Visit supabase.com and create a new project
  2. Locate the SQL file in the repository's supabase folder
  3. Copy its contents
  4. Go to your Supabase project's SQL Editor, paste the code, and run it to create the necessary tables
  5. Once successful, retrieve your Supabase project URL, anon public key, and service role key from the API section
  6. Paste these values into the configuration dashboard
7

Save Configuration

Click "Save All Changes" to ensure all keys and settings are stored.

8

Start the Chatbot

Run the development server again to launch the local server:

npm run dev

Access your chatbot on localhost:3000.

Sign up using your email and confirm the authentication via the link sent to your email.

9

Upload Documents

Go to the dashboard's document processing section.

Upload a file (e.g., a PDF) to test ingestion into the RAG pipeline.

The system processes and chunks the document, storing data in the vector database.

10

Test Your Chatbot

Once a document is uploaded and processed, use the ChatRAG chat interface to query content stored in your vector database.

Try asking questions about the uploaded document to verify the RAG system is working correctly.

11

Basic Customization

Use the visual dashboard to modify:

  • Product name and branding
  • Greeting message
  • Color theme
  • Language settings (support for multiple languages)
12

Next Steps

Explore integrations and advanced features using the dashboard and the codebase:

  • WhatsApp integration
  • Additional AI providers
  • Advanced customization options
  • Deployment to production (Vercel, Fly.io, or other cloud platforms)