QNAP introduces Qsirch AI Mode for semantic file searching

QNAP has announced a major upgrade to its core search application, Qsirch, with the introduction of ‘Qsirch AI Mode’. This new functionality utilizes Vision Language Models (VLM) and Large Language Models (LLM) to enable semantic searching across various file types using natural language.

For documents such as PDF, DOCX, and PPTX, the system understands contextual meaning rather than relying solely on keyword matching. Image searches are enhanced through local VLM inference, allowing users to find photos by describing visual scenes without the need for manual tagging or pre-made albums. For video and audio files, the mode introduces two-way synchronization between transcripts and timelines, enabling users to jump to specific moments by clicking on dialogue text.

A key feature of this update is that all LLM and VLM processing occurs locally on the QNAP NAS device. This local inference approach is designed to maintain data privacy, protect sensitive documents, and eliminate cloud transfer costs, making it suitable for highly regulated or isolated corporate environments.