QNAP introduces RAG Search Beta, a new feature in the Qsirch 5.6.0 search engine that leverages cloud LLMs and RAG technology.
The new RAG Search feature is designed to take traditional search functions on NAS servers to the next level. By using Retrieval-Augmented Generation in combination with cloud-based LLMs, users can search their stored data in a much more intuitive and targeted way. It promises to go beyond conventional file and image searches and understands the context of the data.
Context-aware search capabilities help find files, retrieve information, summarize complex data, and make informed decisions. The system recognizes the user’s intent and adapts search queries accordingly for accurate and relevant results.
Users can choose from various LLMs, including OpenAI ChatGPT, Google Gemini, and Microsoft Azure OpenAI. This offers flexibility in integrating AI search functions according to specific needs.
Data control and ease of use
An important feature of RAG Search is the ability to select specific NAS folders for searches. The system only uploads the most relevant content to the cloud LLM for analysis, which increases accuracy and enhances data control. Users can also determine which file formats are included in the search results.
The feature supports various formats, including Word, Excel, PowerPoint, PDF, TXT, and emails (.eml). Search results can be linked to up to five relevant documents, which increases data validity and deepens analysis. All results reflect the most recent versions of files on the NAS, ensuring that information is always up to date.