Rapport
Rapport is a simple python LLM chatbot interface. It uses Ollama to run LLMs locally and Streamlit to create the chatbot user interface.
I started building this because I was excited by the idea that one could (as of Jan 2025) run pretty competent LLM models on a laptop. While this app was originally just a learning exercise, I built enough bits that I’m using it day to day. I also built out connections to a few remote providers so I could try out frontier models.
In the end, I built out a reasonably functional chat interface with quite a lot of features. In the end, I found the development teams at Anthropic and OpenAI were more than a match for my hobbyist build, but I still flip between Claude in Anthropic’s UI and Claude in Rapport, particularly as I can better tune the prompt in Rapport.
