Show HN: How to Use LLMs (github.com)

A new tool called Sidelong has been introduced for utilizing Large Language Models (LLMs) in a unique chat interface. This application features a sideways layout where each question-and-answer pair appears as a separate page, enabling users to navigate through conversations visually, with topics color-coded for easy identification. The tool operates through the Claude Code CLI, integrating seamlessly with a Claude subscription, and utilizes local embeddings via the Ollama model, ensuring data privacy by keeping all embeddings on the user's machine. Sidelong's innovative design aims to enhance user experience by allowing for easy tracking of conversation threads, as follow-up questions share color with their related topics based on classifier confidence. With a focus on performance, the system continuously recalibrates conversation placements based on real-time information, providing a dynamic interaction model. Moreover, being built using plain JavaScript, it maintains minimal dependencies, making it lightweight and accessible. This tool represents a significant step in making LLMs more interactive and user-friendly, potentially driving wider adoption in both personal and professional contexts by simplifying conversational management.