llcat is a stateless CLI tool that calls OpenAI-compatible /chat/completions and Anthropic Messages API endpoints, also supporting Ollama, OpenRouter, SGLang, vLLM, LocalAI, and llama.cpp.

It accepts keys, models, prompts, and server URLs as command-line arguments using plain strings, file references with line numbers, and jq syntax. The tool enables parallel query execution across multiple models and servers without persistent state, and is distributed on PyPI under the MIT license.