🔷 DeepFileX

DeepFileX - Ultra-Lightweight Desktop File Contents Search & Analysis Solution

Latest: v3.3.0 (2026-07-21)

---

🎯 Overview

DeepFileX is a high-performance desktop search utility written in Rust. It is designed to scan, index, and retrieve not only filenames but also the full-text content of documents across your local storage.

By building a secure local database index using parallel multi-threaded scanning and SQLite FTS5, DeepFileX allows you to search through the actual contents of PDFs, HWP/HWPX, Office documents, code files, and text files in milliseconds—completely offline, ensuring 100% data privacy.

---

🚀 Quick Start

Build & Run from Source

1. Ensure you have the Rust toolchain installed.
2. Navigate to the hakar-core directory and run:

   # Run in developer mode
   cargo run

3. To compile the optimized, standalone release binary:

   # Compile optimized release binary
   cargo build --release

4. Find the standalone executable at: target/release/DeepFileX.exe

---

⭐ Core Features

🔒 Real-Time Blackbox JSON Diagnostic Logger

• Append-Only Diagnostic Log: Appends all user & system events in real-time to %USERPROFILE%\Documents\DeepFileX\Logs\blackbox_log.json without overwriting historical logs.
• Rich Diagnostic Metadata: Captures unique UUID session_id, datetime timestamps (YYYY-MM-DD HH:MM:SS.mmm), and formatted execution duration ("145ms", "5.981s").
• Lifecycle Tracking: Automatically logs application startup and shutdown events (eframe::App::on_exit).

⚡ 2-Phase Engine Architecture & Performance

• Phase 1 (MFT Filename Scan): Reads 100% of all file and folder entries (including .mp4, .exe, .zip, .pdf, .hwp) for Everything-level sub-millisecond filename searching.
• Phase 2 (Content & AI Dispatcher): Early bypass (100% Skip) for non-document files before reaching heavy parsing pipelines, protecting CPU & RAM footprint by >90%.
• B-Tree SQL Query Optimization: High-speed indexing queries without function wrapper overhead (REPLACE(LOWER(...))), cutting query response time to sub-millisecond speeds.

📄 Native High-Performance Parsers

• Native HWP / HWPX Parser: Built-in OLE2, Zstd, and XML text extraction natively compiled into core engine without requiring external DLL plugins.
• Embedded Pdfium Engine: Integrates Google Chromium's official Pdfium parser via static bundling (include_bytes!), achieving 99% accurate PDF text extraction.
• SQLite FTS5 + Compression: Ultra-fast content lookup paired with Zstd Lv3 text compression for minimal disk footprint.

---

📁 Supported File Formats & Search Level

DeepFileX employs a hybrid indexing strategy: dedicated high-performance parsers for primary document types, and an optimized binary strings extractor as a universal fallback.

Category	Extensions	Search Level / Implementation
📄 PDF Documents	.pdf	Full-Text (Dedicated Embedded Pdfium)
🇰🇷 HWP / HWPX	.hwp, .hwpx	Full-Text (Native Core OLE2 / Zstd / XML Engine)
📁 Word / Excel / PPT	.docx, .xlsx, .pptx	Full-Text (Dedicated XML/Zip Parsers)
⚙️ Configs & Scripts	.yaml, .yml, .ini, .cfg, .conf, .toml, .env, .properties, .txt, .csv, .log, .srt, .vtt, .md, .json, .xml	Full-Text (Dedicated Text Parser)
💻 Code Files	.py, .js, .java, .c, .cpp, .h, .cs, .rs, .go, .sh, .bat, .html, .css	Full-Text (Dedicated Text Parser)
🔌 Dynamic Plugins	.dwg, .dxf	Full-Text (Optional Dynamic CAD DLL Plugin via 🔌 Plugins Manager)
🖼️ Media & Archives	.zip, .png, .jpg, .mp4, .avi, .dat, .bin, .dll, .exe, etc.	Filename Search Only (Phase 2 Early Bypass)

---

💾 System Requirements

• OS: Windows 10 / 11 (64-bit)
• RAM: 4GB+ recommended
• Storage: Less than 10MB (Single binary size: ~4.5MB)
• Rust Toolchain: 1.75+ (Developer compilation only)

---

📊 Performance Metrics

• First Scan Speed (MFT): Under 5 seconds for 1.8 Million files
• Search Response Time: Sub-millisecond (~0.1ms)
• Idle Memory Footprint: Under 10MB
• Active Scan Memory Footprint: Under 30MB
• Standalone Binary Size: ~4.5MB (LTO & Strip optimized)

---

📈 Recent Updates

v3.3.0 (2026-07-21)

• 🔒 Real-Time Blackbox JSON Logger: Added append-only logging to %USERPROFILE%/Documents/DeepFileX/Logs/blackbox_log.json with UUID session tracking, datetime, duration, and 100% English GUI.
• ⚡ 2-Phase Engine Architecture: Implemented Phase 1 MFT 100% filename scan + Phase 2 Content/AI Dispatcher 100% early skip for non-document files.
• 🚀 B-Tree SQL Query Optimization: Removed function wrappers (REPLACE(LOWER(...))) on external DB queries for instant B-Tree search responses.
• 🔍 Smart Loading Indicator: Display Searching index databases... loading spinner only after typing non-empty search queries before results render.
• ⚙️ Dedicated Update Check Dialog: Separated ⚙️ Settings... from 🔄 Check for Updates in top menu.
• 🔌 Streamlined Plugin Manager: Removed HWP/HWPX from Plugin Manager UI since HWP/HWPX parsing is 100% natively built into core engine.

📖 See CHANGELOG.md for complete historical details.

---

🛠️ Development & Build Guide

Directory Structure

hakar-core/
├── src/                  # Rust source code
│   ├── main.rs           # GUI Application Entry (egui)
│   ├── parser.rs         # File text extractors (Pdfium, HWP, zip, xml)
│   ├── db.rs             # SQLite FTS5 database mapping
│   ├── ntfs.rs           # NTFS MFT parser
│   ├── blackbox/         # Real-time JSON Blackbox Logger subsystem
│   └── update/           # Auto-updater subsystem
├── Cargo.toml            # Dependencies and Release profiles
├── build.rs              # Windows resource compiler (icon mapping)
└── README.md             # Project documentation

Build Command

To compile the production-ready optimized executable, run:

cargo build --release

The output binary will be generated at target/release/DeepFileX.exe.

---

🤝 Contributing

This is a proprietary utility belonging to HAKAR. External contributions require signing an NDA.