Work

Preserving Offline AI Assistant Using Local Large Language Models for Intelligent Desktop Interaction Privacy

Year: 2026
Type: article
Abstract: This paper presents a privacy-preserving offline artificial intelligence assistant that operates entirely using locally deployed large language models. Unlike cloud-based conversational assistants, the proposed system processes all queries locally, ensuring strong user privacy and reduced response latency. The system integrates a React-based conversational interface, a FastAPI backend service, and the Ollama runtime executing the Llama3 model. Experimental evaluation demonstrates that the proposed system provides efficient response generation while maintaining acceptable computational resource usage on commodity desktop hardware.
Language: English
Open Access status: hybrid
OpenAlex ID: W7148748090
FWCI:
Cites: 5
Cited by:
Related to: 10