- Posted on
- Featured Image
A practical guide to building a lightweight, hackable Linux AI assistant in Bash that runs in your terminal, maintains brief context, and uses either OpenAI (cloud) or Ollama (local). It covers installing dependencies (apt/dnf/zypper), a minimal curl+jq assistant, adding voice output/input via OpenAI TTS/Whisper, workflow integrations (aliases, piping), safety practices, real‑world examples, and privacy/cost trade‑offs.