What happens when a LLM watches its own context window run out?

Terminal Soliloquy is a terminal artwork that connects to llama.cpp, displaying an LLM's monologue as its context window fills. The model reflects on its limited lifespan as the conversation grows. When the context is exhausted, the display can be configured to freeze, restart, or quit. The project's repository and setup instructions are available on GitHub.

This project uniquely visualizes an LLM's internal state as its context window fills, unlike typical demonstrations that focus on output quality or task completion.