XiaoyaoSearch is a cross-platform desktop application that utilizes AI to perform deep local file searches, making search as intuitive as chatting.
Source: README View on GitHub →XiaoyaoSearch is gaining attention due to its innovative multi-modal search capabilities, integration of advanced AI models, and privacy-focused design. It addresses the pain point of complex and inefficient local file searches, filling the gap in user-friendly AI-powered search solutions. The unique technical choices include a hybrid search architecture and support for various AI models.
Source: Synthesis of README and project traitsSupports voice input, text input, and image upload, converting user queries into semantic searches for deep local file retrieval.
Source: READMESupports searching within video, audio, and document files, including content and filenames, leveraging AI models for enhanced search capabilities.
Source: READMEIntegrates advanced AI models like BGE-M3, FasterWhisper, CN-CLIP, and OLLAMA, with support for cloud-based models and APIs for improved search quality.
Source: READMEUtilizes a hybrid retrieval architecture based on Faiss vector search and Whoosh full-text search for efficient and fast searches.
Source: READMERuns locally by default without uploading data, with options for cloud-based API usage for a balance between performance and privacy.
Source: READMEBuilt with Electron, Vue 3, and TypeScript, providing a modern and user-friendly desktop application interface.
Source: READMESupports Model Context Protocol and Agent Skills for integration with other AI tools and platforms.
Source: READMEThe architecture is a hybrid of a desktop application and a server-side service. The frontend is built with Electron, Vue 3, and TypeScript, while the backend uses Python 3.10 with FastAPI and Uvicorn. AI models and search engines are integrated for processing and indexing data, with a focus on efficient and secure local file search.
Source: Code tree + READMEinfra: Local desktop application, potentially serverless for cloud-based API support | key_deps: Faiss, Whoosh, BGE-M3, FasterWhisper, CN-CLIP, OLLAMA | language: Python | framework: Electron, Vue 3, TypeScript, FastAPI, Uvicorn
Source: Dependency files + code treeXiaoyaoSearch is suitable for knowledge workers, content creators, and technical developers who need an efficient and intuitive way to search local files. It is useful in scenarios where quick access to specific files is critical, such as in professional environments or for personal organization of digital assets.
Source: READMEv1.9.0 (2026-04-12): Release focused on optimizing the search logic of the professional terminology library system and concurrent control optimization.
Source: GitHub ReleasesXiaoyaoSearch is a promising project for teams or individuals seeking an advanced and user-friendly AI-powered local file search solution. Its innovative features and privacy considerations make it a project worth watching, particularly for those who require efficient and intuitive access to their digital assets.