Hybrid Semantic & BM25 Search
Understands technical jargon, synonyms, and natural language phrasing. Returns exact passages even when documents use different terms from your search query.
AI SEMANTIC SEARCH & VIEWER
Find exact answers inside manuals, contracts & guides.
Ask in plain language. Semantic search helps find relevant passages even when documents use different wording, while OCR extracts text from scanned pages.
AI Direct Answer
Gemini 3.8 Flash
Synthesizing answer from your documents...
Sources cited:
🛡️ SEARCH RESULTS (100% Local Mode)
MY LIBRARY
My Documents
Upload Document
PDF, Word, text, or image. Up to 60 MB.
or drag and drop files here
Text is automatically extracted & indexed for semantic search.Your library is empty. Upload your first manual or document above.
AI DOCUMENT INTELLIGENCE
DocSearch – AI Semantic Document Search & Intelligent Reader
Standard keyword search (Ctrl+F) fails when wording differs. DocSearch combines multilingual dense vector embeddings, BM25 keyword ranking, and Tesseract OCR to find exact answers inside equipment manuals, contracts, technical specifications, and scanned PDFs in seconds.
Scanned PDF & Image OCR
Built-in optical character recognition for scanned manuals and diagrams in English, Finnish, and Spanish. Preserves headings, tables, and page numbers.
Direct AI Answers with Gemini
Unlock Google Gemini AI mode with your own API key to get instant, synthesized answers cited directly from your documents with page-level badges.
Private & Session-Isolated Storage
Your documents remain in your private session library. Uploaded materials are never used for public AI model training or shared with third parties.
Frequently Asked Questions
How does DocSearch differ from standard PDF search?
Regular search tools match exact characters only. DocSearch understands query intent and technical semantics, pinpointing the right paragraphs, tables, and sections.
What file formats can be uploaded and searched?
DocSearch supports PDF, Microsoft Word (.docx), plain text (.txt), and images (.png, .jpg, .jpeg, .tif, .tiff).
How does it handle scanned paper documents?
Pages without selectable text are automatically processed using high-accuracy Tesseract OCR with Finnish, English, and Spanish language models.
Who is DocSearch built for?
Engineers, property managers, electricians, HVAC technicians, maintenance specialists, architects, quality auditors, and procurement professionals.