Local Brain Document Search

Local Brain Document Search is an offline semantic search engine and RAG (Retrieval-Augmented Generation) provider for Android. It allows you to find exact information in your local files by meaning and context, moving beyond traditional keyword matching. All artificial intelligence processing runs strictly on-device to ensure zero data leaks and absolute privacy. 🎯 ADVANCED SEMANTIC SEARCH • On-Device AI Engine: Uses quantized MiniLM embedding models (Fast or Deep profiles) to understand search intent. • Hybrid Search: Combines AI semantic similarity with exact keyword matching for highly accurate results. • Multi-Format Support: Natively indexes PDF, DOCX, TXT, Markdown, and CSV files. • Built-in OCR: Extracts and searches text within scanned PDFs and images using ML Kit (offline Latin alphabet recognition). 🔒 TOTAL PRIVACY & OFFLINE MODE • 100% Local: Your documents and the vector index are never uploaded to any cloud server. • No Invasive Permissions: Uses the native Android Storage Access Framework (SAF). It only accesses the specific folders you choose with read-only permissions. It never modifies your files. • Minimal Connection: Core search and AI features are completely offline. Internet access is only used to display advertisements. ⚙️ INTEGRATION FOR DEVELOPERS & AI AGENTS (RAG) • Local RAG API: Allows other apps and AI agents to query your indexed documents safely via Deep Links (rag://api/...) and a read-only Content Provider. • Strict Access Control: Secure your data using verified app whitelists, secret tokens, or manual request approvals. • Developer Tools: Features an internal playground to test endpoints and a detailed local access log. 🔋 PERFORMANCE & BATTERY EFFICIENCY • Incremental Indexing: Only processes new or modified files to conserve system resources. • Background Sync: Scheduled to update every 12 hours. By default, background indexing only runs while the device is charging. • Full Control: Visible execution via a foreground service with a progress notification and an immediate stop button. Note: The application interface is in English. The semantic search engine is highly optimized for processing and finding content in Spanish and English.

Disfruta la app, agradecemos las sugerencias.