Local AI Chat: On-Device LLM

Offline LLM Assistant is a local-first AI assistant for technical notes, code snippets, summaries, rewrites, and quick Q&A.

The app runs generation on your device after you install or import a supported Qwen GGUF model. Prompts and attached context are not sent to an AI server for inference. You can choose a fast model for mid-range devices or an optional quality model for higher-RAM devices.

Key features:
- On-device Qwen GGUF model support.
- Terminal-style transcript with copyable assistant output.
- Code explanation, small function generation, rewrite, summary, and Q&A modes.
- Optional context buffer for local documents, logs, and snippets.
- User-submitted output reporting for offensive or unsafe AI output.

Model download disclosure:
- The app can download large model files from Hugging Face.
- The fast model is about 1.28 GB.
- The quality model is about 2.5 GB and is recommended for 8 GB+ RAM devices.
- Inference remains on-device after a model is installed.

Important limitations:
- Small local models can be wrong or incomplete.
- Do not rely on generated output as professional legal, medical, financial, or security advice.
- Report offensive or unsafe output from inside the app