This is an LLM-agent that acts as a travel assistent. The application provides a chat window and a document editor. The user can interact with the chat window to chat with the assistant and can use the document editor to create a travel document, containing all information related to an upcoming trip. Based on the chat history, the LLM-agent updates (adds, removes, changes) the document. There are 1. event-based updates: the user chats with the assistant, and 2. scheduled updates: the assistant checks if the info in the document is up to date.
- Run the React frontend
cd gpta-frontend
npm start
- Start the LangChain agent
cd agent-backend
python app.py
Alternative, start the Open AI inference server
cd gpta-backend
node server.js
Use the application on http://localhost:3000
This repository contains three folders in its root directory:
-
agent-backend: This folder contains a Flask app that functions as a REST-api for the application. The app contains a LangChain agent with GPT-4 as its core.
-
gpta-backend: This folder contains an Express.js server, running on http://localhost:5001. The server forwards user inputs via the Open AI api and returns the response without any intermediate processing.
-
gpta-frontend: This folder contains a React app that functions as the frontend of the application.