qqqa

Why we included this project

Most LLM tools want to own a chat window. qqqa takes the opposite route: it is a stateless CLI where qq answers a single question in one shot and qa acts as a one-step agent that can read a file, write a file, or run a command, always asking for confirmation before anything executes. Because every invocation is independent, the tool composes cleanly with pipes and scripts, and qq stays read-only while qa gates its tools behind a prompt. Out of the box it has profiles for OpenRouter, OpenAI, Groq, Gemini, a local Ollama runtime, and the Codex and Claude Code CLIs, so you can reuse a subscription you already pay for or point it at a local model. For anyone who wants quick, repeatable answers and small delegated shell tasks without leaving the terminal, the setup is minimal and the safety rails are reassuring.