LLM

LLM is an open-source CLI and Python library by Simon Willison for prompting large language models from the terminal. It works with hosted models from OpenAI, Anthropic, Google and others, and with models installed on your own machine through plugins.

You can run one-off prompts or chats, attach images and files, use system prompts and templates, and extract structured data against a schema. Every prompt and response is logged to a SQLite database, which makes past work easy to search and analyze.

LLM can also generate and store embeddings, give models tools to execute, and point at any OpenAI-compatible endpoint such as LM Studio. A plugin system adds new model providers and embedding models. It installs with pip, Homebrew, pipx or uv.