chrisryugj/Docufinder

Anything — 파일 이름이 아니라 본문으로 찾는 문서 검색기. HWP·PDF·Office 전문 검색을 오프라인으로. Windows·macOS | Offline content search across Korean documents (HWP, PDF, Office). Windows and macOS

What it solves

It solves the difficulty of finding specific information buried within large volumes of local files. Instead of searching only by filename, users can search by the actual content inside various document formats, including complex Korean HWP files, PDFs, and even images via OCR.

How it works

The application indexes local folders to enable lightning-fast keyword and semantic searches. It uses a hybrid approach combining SQLite FTS5 for keyword matching and HNSW-based vector search for semantic similarity. For AI features like natural language Q&A and summarization, it uses RAG (Retrieval-Augmented Generation) by sending only relevant text chunks to compatible LLM providers (like Gemini or OpenAI-compatible servers).

Who it’s for

  • Individuals looking for a powerful, privacy-focused local search tool.
  • Professionals working with sensitive data who need to keep their document indexing entirely offline.
  • Users who need to search through diverse file types including legacy HWP formats and scanned images.

Highlights

  • Hybrid Search: Combines traditional keyword matching with semantic vector search.
  • Local-First Privacy: All indexing, embedding, and OCR happen locally; AI features can be completely disabled to ensure zero network communication.
  • Extensive Format Support: Supports HWP/HWPX, DOCX, XLSX, PDF, and images via OCR.
  • AI Capabilities: Offers natural language Q&A (with source citations) and document summarization via external LLM APIs.
  • Advanced Preview: Includes a built-in viewer that preserves original layouts for various document types.