Skip to content

An LLM powered chatbot that can answer questions based on your specific data

Notifications You must be signed in to change notification settings

Zeusmist/langchain-powered-ai-chatbot

Repository files navigation

Langchain Powered AI Chatbot (Scrimba Tutorial)

An LLM powered chatbot that can answer questions based on your specific data. This project is based on the Scrimba Tutorial by Tom Chant.

App Flow

App Flow

Knowledge Store

Knowledge Store

Getting Started

Prerequisites:

  • Supabase Account
  • Google AI Studio Account (This project uses Google AI not OpenAI)
  • Run the query found in match_documents.sql in your Supabase database

Install the dependencies:

yarn install

Set the environment variables:

cp .env.example .env

# open .env and modify the environment variables

Run the app

yarn start

Improving Performance

  • Update the chunk size and overlap size used by the text splitter
  • Test different number of chunks retrived
  • Prompt engineering
  • Try different LLM settings based on your use case