GitHub page: https://github.com/wch/llm-chat

This project contains two Shiny chat applications:

• basic-chat provides a chat interface to an LLM from OpenAI, similar to the one available at chatgpt.com.
• rag-chat provides an example of Retrieval-Augmented Generation, where the chat application queries documentation from quarto.org Quarto to look for relevant pieces of information, then passes that information to an LLM, and asks the LLM to answer the user query using that information.

To run these apps, first install some packages from R:

if (!require("pak")) install.packages("pak")  # Instal pak if not already installed
pak::pak(c("dotenv", "shiny", "hadley/elmer", "jcheng5/shinychat"))

Then create a file named env in the project directory with API keys that you have been provided. (Replace the XXXXXXXX below with the actual API keys.)

OPENAI_API_KEY="XXXXXXX"
ANTHROPIC_API_KEY="XXXXXXX"
EMBEDDING_IO_API_KEY="XXXXXXXX"

Next, open basic-chat/app.R in the editor and click on the Run button. You should now have a running chat application!

For the RAG app, open rag-chat/app.R and click on the Run button.