LLM CLI

LLM is a command-line utility and Python library by Simon Willison that gives you one consistent way to talk to large language models — OpenAI, Anthropic’s Claude, Google’s Gemini, and dozens more, including models running on your own machine. The open-source tool has earned more than 12,600 stars on GitHub, is licensed under Apache-2.0, and was actively developed in 2026.

Overview

LLM treats the shell as a first-class home for language models. Pipe a file in and ask questions about it, run one-off prompts, hold multi-turn chats, extract text from images, or generate embeddings — all with short, scriptable commands. Every interaction is logged to a SQLite database, so your prompt history becomes searchable data you can analyze with tools like Datasette. A rich plugin system adds new model providers (including local ones via Ollama) and new capabilities without changing the core interface, and the same features are exposed as a Python API for use inside your own programs.

Key Features

• One CLI for OpenAI, Claude, Gemini, Qwen, DeepSeek, Mistral and many other providers
• Prompt files, pipe input, and continue or start interactive chats from the terminal
• Automatic logging of prompts and responses to SQLite for later analysis
• Embeddings support for building search and clustering workflows
• Plugin ecosystem for additional hosted and local models
• Usable as a Python library as well as a standalone command-line tool
• Installable via pip, Homebrew, pipx or uv

💳 Pricing

LLM itself is free and open source under the Apache-2.0 license. You pay only for any hosted model APIs you choose to call with your own keys; local models cost nothing beyond your hardware.

Who Is It For

LLM is perfect for developers, data journalists, researchers and shell power-users who want fast, scriptable access to many models at once — and a permanent, queryable record of everything they have asked.