Bull Bot – a chatbot designed to comprehend natural language queries and respond with precise answers, accompanied by the most relevant website links. Live at BullBot.tech
- Vite.js / Node.js / Python / langchain /
- Database: Pinecone
- Embedding Model: BAAI/bge-small-en
- Large Language Model: OpenAI API
- Deployment: Vercel / Cyclic / AWS EC2
- The client hosted on Vercel is a web app that allows users to ask questions and get answers from the server.
- The server hosted on Cyclic, which is an Express app, makes request to the Flask server to get answers to questions.
- The Flask server hosted on AWS EC2 hosts the embedding model and makes calls to OpenAI API to generate answers to questions.
- The dataset is automatically generated and uploaded to Pinecone database by the dataset pipeline.
Client server is a web app that allows users to ask questions and get answers from the server.
Express server get requests from the client and make calls to the Flask server to get answers to questions.
Server hosts the embedding model and make calls to OpenAI API to generate answers to questions.
Dataset Pipeline is pineline to generate dataset by srapping raw data from the University of South Florida website, processing the raw data to generate the final dataset, and uploading the dataset to Pinecone database.