Releases: 5ymph0en1x/DLER
Releases · 5ymph0en1x/DLER
v1.2.1
Adaptive Connection Management
High-Performance NZB Downloader for Windows
This release introduces intelligent connection management with adaptive pipelining and improved retry logic.
What's New in v1.2.1
Adaptive NNTP Pipelining
- RTT-based pipeline depth - Automatically adjusts pipeline depth based on measured round-trip time
- Smart scaling - Low latency (<25ms) = 25 requests, medium (~50ms) = 40 requests, high (>100ms) = 60+ requests
- Real-time display - Progress bar now shows pipeline depth and average RTT (e.g.,
P:40@25ms)
Connection Resilience
- Exponential backoff with jitter - AWS-style retry logic prevents thundering herd on reconnection
- Connection health monitoring - Tracks consecutive failures, success rate, and response times
- Dynamic connection pool - Health-based connection rotation for improved reliability
Visual Improvements
- Matrix-style banner - New animated SVG banner with falling green symbols for README
Download
| Edition | Size | Description |
|---|---|---|
DLER_v1.2.1.exe |
~1 GB | Ultimate - GPU-accelerated PAR2 repair (CUDA) |
DLER_Basic_v1.2.1.exe |
~38 MB | Basic - CPU-only, lightweight |
Both editions are functionally identical. The only difference is GPU acceleration for PAR2 repair.
System Requirements
- OS: Windows 10/11 (64-bit)
- CPU: x64 with AVX2 support (Intel Haswell 2013+ / AMD Excavator 2015+)
- RAM: 4 GB minimum, 32+ GB recommended for RAM mode
- GPU (Ultimate only): NVIDIA GPU with CUDA 12.x or 13.x
Quick Start
- Download the appropriate edition
- Launch and click Settings (gear icon)
- Configure your Usenet server (host, port, SSL, credentials)
- Set download and extraction directories
- Drag & drop NZB files to start downloading
Settings are saved to ~/.dler/config.json
Changelog
v1.2.1 (2026-01-17)
- New: Adaptive NNTP pipelining based on RTT measurement
- New: Exponential backoff with jitter for connection retry
- New: Connection health monitoring and dynamic pool management
- New: Matrix-style animated banner for README
- Improved: Progress display shows pipeline depth and RTT
v1.2.0 (2026-01-16)
- Fixed: Multi-volume RAR extraction with encrypted headers now works correctly
- Fixed: Password from NZB metadata properly used for all extraction stages
- Fixed: Duplicate folder issue (Release.Name/Release.Name/) automatically corrected
- Improved: Pre-sorted volume order for encrypted archives (uses NZB buffer index)
- Improved: Edition detection system (Basic vs Ultimate)
- Improved: GPU options properly disabled in Basic Edition
Previous Releases
- v1.1.2 - Incremental PAR2 verification, early damage detection
- v1.1.1 - Dual edition system, CUDA kernel warmup
- v1.1.0 - RAM mode, GPU acceleration, Windows long path support
- v1.0.0 - Initial release
Bundled Tools
| Tool | Version | License |
|---|---|---|
| 7-Zip | 23.01 | LGPL |
| par2j64 | 1.3.2.6 | GPL |
| UnRAR | 7.0 | Freeware |
License
MIT License - See LICENSE file for details
Download at wire speed.
v1.2.0
What's New in v1.2.0
Encrypted Archive Fixes
- Multi-volume RAR with encrypted headers - Now works correctly with password-protected archives where headers are also encrypted
- Password propagation - Password from NZB metadata (
<meta type="password">) is now properly passed to all extraction stages - Pre-sorted volume order - Uses NZB buffer index for correct volume ordering when headers can't be read
Extraction Improvements
- Duplicate folder fix - Automatically corrects
Release.Name/Release.Name/nested folder issue - Better error reporting - CUDA detection errors now displayed in Settings for easier debugging
Build System
- Edition detection - Basic and Ultimate editions now properly identified via runtime hooks
- GPU options grayed out - Basic Edition clearly shows GPU options as unavailable
- Improved CuPy bundling - Fixed CUDA module collection for PyInstaller builds
Download
| Edition | Size | Description |
|---|---|---|
DLER_v1.2.0.exe |
~1 GB | Ultimate - GPU-accelerated PAR2 repair (CUDA) |
DLER_Basic_v1.2.0.exe |
~38 MB | Basic - CPU-only, lightweight |
Both editions are functionally identical. The only difference is GPU acceleration for PAR2 repair.
System Requirements
- OS: Windows 10/11 (64-bit)
- CPU: x64 with AVX2 support (Intel Haswell 2013+ / AMD Excavator 2015+)
- RAM: 4 GB minimum, 32+ GB recommended for RAM mode
- GPU (Ultimate only): NVIDIA GPU with CUDA 12.x or 13.x
Quick Start
- Download the appropriate edition
- Launch and click Settings (gear icon)
- Configure your Usenet server (host, port, SSL, credentials)
- Set download and extraction directories
- Drag & drop NZB files to start downloading
Settings are saved to ~/.dler/config.json
Changelog
v1.2.0 (2026-01-16)
- Fixed: Multi-volume RAR extraction with encrypted headers now works correctly
- Fixed: Password from NZB metadata properly used for all extraction stages
- Fixed: Duplicate folder issue (Release.Name/Release.Name/) automatically corrected
- Improved: Pre-sorted volume order for encrypted archives (uses NZB buffer index)
- Improved: Edition detection system (Basic vs Ultimate)
- Improved: GPU options properly disabled in Basic Edition
Previous Releases
- v1.1.2 - Incremental PAR2 verification, early damage detection
- v1.1.1 - Dual edition system, CUDA kernel warmup
- v1.1.0 - RAM mode, GPU acceleration, Windows long path support
- v1.0.0 - Initial release
Bundled Tools
| Tool | Version | License |
|---|---|---|
| 7-Zip | 23.01 | LGPL |
| par2j64 | 1.3.2.6 | GPL |
| UnRAR | 7.0 | Freeware |
License
MIT License - See LICENSE file for details
Download at wire speed.
v1.1.2
🚀 What's New
Incremental PAR2 Verification
DLER now verifies files as they complete downloading, rather than waiting for the entire download to finish. This means:
- Early damage detection - Know if repair is needed before the download completes
- Per-file MD5 verification - Each file is checksummed immediately upon completion
- Parallel verification - Multiple files verified simultaneously without blocking downloads
- Faster feedback - See verification status in real-time in the logs
Post-Processing Optimization
- Improved pipeline efficiency for archive verification
- Better integration between download engine and verification system
- Reduced latency between download completion and post-processing start
📦 Downloads
| Edition | File | Size | Description |
|---|---|---|---|
| Ultimate | DLER_v1.1.2.exe |
~1 GB | Full version with CUDA/GPU acceleration |
| Basic | DLER_Basic_v1.1.2.exe |
~37 MB | Lightweight CPU-only version |
Which one should I download?
- DLER Basic - Recommended for most users. Smaller download, full functionality, CPU-based PAR2 repair.
- DLER Ultimate - For users with NVIDIA GPUs who want GPU-accelerated PAR2 repair (~15 GB/s on RTX 4090).
Both editions include the AVX2-accelerated yEnc decoder (3+ GB/s) and all core features.
💻 System Requirements
- OS: Windows 10/11 (64-bit)
- CPU: x86-64 with AVX2 support (Intel Haswell 2013+ / AMD Excavator 2015+)
- RAM: 2 GB minimum, 32+ GB recommended for RAM mode
- GPU (Ultimate only): NVIDIA GPU with CUDA 13.x
🔄 Changelog
Added
- Incremental PAR2 verification - verifies files as they complete downloading
- Early damage detection - know if repair is needed before download finishes
- Per-file MD5 verification with parallel computation
Improved
- Post-processing pipeline optimization
- Better resource management during verification
📝 Notes
- No installation required - standalone executable
- Configuration stored in
~/.dler/config.json - Bundled tools: 7-Zip 23.01, par2j64 (MultiPar 1.3.2.6)
v1.1.1
Dual Release: Basic & Ultimate
Two editions to match your hardware. Same features, different GPU support.
What's New in v1.1.1
Bug Fixes
- Fixed filename parsing - Files no longer have
(1/0)suffix artifacts in their names - Fixed multi-download sessions - You can now queue multiple NZBs without restarting the app
- Automatic reconnection - Connections are refreshed between downloads to prevent stale socket errors
Improvements
- CuPy kernel warmup - GPU kernels are pre-compiled at startup, eliminating first-use latency
- VRAM management - Memory is properly released after PAR2 repair operations
- Better error handling - VRAM cleanup even on failures
Choose Your Edition
| DLER Basic | DLER Ultimate | |
|---|---|---|
| Size | ~38 MB | ~1 GB |
| RAM Mode | ✅ | ✅ |
| yEnc Decoding | ✅ 3+ GB/s (AVX2) | ✅ 3+ GB/s (AVX2) |
| PAR2 Verification | ✅ CPU | ✅ CPU |
| PAR2 GPU Repair | ❌ | ✅ CUDA |
| Requirements | Any modern CPU | NVIDIA GPU + CUDA 13.x |
Which one should I download?
Download Basic if:
- You don't have an NVIDIA GPU
- You have limited disk space
- You rarely need PAR2 repair (most downloads complete without errors)
- You want the smallest possible download
Download Ultimate if:
- You have an NVIDIA GPU (GTX 1060+, RTX series)
- You want GPU-accelerated PAR2 repair
- You frequently download from unreliable sources requiring repair
- Disk space is not a concern
Note: Both editions are identical in functionality. The only difference is GPU acceleration for PAR2 repair operations. If a download completes without errors, you won't notice any difference.
Downloads
| Edition | Filename | Size |
|---|---|---|
| Basic | DLER_Basic.exe |
~38 MB |
| Ultimate | DLER.exe |
~1 GB |
System Requirements
Basic Edition
| Component | Minimum |
|---|---|
| OS | Windows 10 64-bit |
| CPU | x86-64 with AVX2 (Intel Haswell+ / AMD Zen+) |
| RAM | 2 GB (32+ GB for RAM mode) |
Ultimate Edition
| Component | Minimum |
|---|---|
| OS | Windows 10 64-bit |
| CPU | x86-64 with AVX2 |
| RAM | 2 GB (32+ GB for RAM mode) |
| GPU | NVIDIA with CUDA 13.x support |
Installation
No installation required. Download the .exe and run it.
First Launch
- Go to Settings tab
- Configure your Usenet server (host, port, SSL, credentials)
- Set your download directory
- (Optional) Enable RAM mode if you have 32+ GB RAM
Ultimate Edition - GPU Setup
The Ultimate edition includes CUDA libraries. If you encounter GPU errors:
- Ensure you have an NVIDIA GPU
- Install/update your NVIDIA drivers
- The app will automatically fall back to CPU if GPU is unavailable
Changelog
v1.1.1
- Fixed: Filename artifacts
(1/0)no longer appear in downloaded files - Fixed: Multiple NZBs can now be downloaded in sequence without restart
- Fixed: Automatic reconnection between downloads
- Added: CUDA kernel warmup for reduced first-use latency
- Added: Proper VRAM cleanup after GPU operations
- Added: Dual-edition build system (Basic/Ultimate)
v1.1.0
DLER v1.1.0 - RAM Mode & GPU Acceleration
High-performance NZB downloader engineered for maximum throughput on modern hardware.
What's New
RAM Mode
Full in-memory download and processing pipeline for systems with 32+ GB RAM:
- Zero disk I/O during download - All data buffered in RAM until extraction
- Parallel MD5 verification - ~5 GB/s throughput
- Direct extraction - Only final content written to disk
GPU Acceleration (Optional)
- CUDA/CuPy support for PAR2 matrix operations
- Requires NVIDIA GPU with CUDA 13.x
- Automatic CPU fallback when unavailable
Windows Long Path Support
- Handles paths exceeding 260 characters
- No registry tweaks required
Bug Fixes
- Fixed archive detection in RAM mode
- Fixed file naming with obfuscated NZBs
- Fixed CuPy/NumPy compatibility with Python 3.14
Requirements
| Component | Minimum | Recommended |
|---|---|---|
| OS | Windows 10 (64-bit) | Windows 11 |
| CPU | x86-64 with AVX2 | Ryzen 9 / i9 |
| RAM | 2 GB | 32+ GB (RAM mode) |
| GPU | - | NVIDIA CUDA 13.x |
Installation
Download DLER.exe below - standalone executable, no installation required.
v1.0.4
What's New
Speed Graph Improvements
- Fine line rendering - Cleaner 1px line without glow effect
- Full history - Graph now keeps entire download history, compressing to fit (no more sliding window)
- Better smoothing - Increased splinesteps for smoother curves
- Active only during download - Graph freezes during post-processing to show final speed profile
Direct-to-Destination Optimization
- Smart NZB analysis - Detects if release contains archives (RAR/ZIP/7z) before downloading
- Skip unnecessary copy - For non-archive releases, downloads directly to final destination
- Massive time savings - Eliminates the post-download copy step (e.g., 4 minutes saved on 18GB files)
New LOGS Panel
- Replaced QUEUE with LOGS - Real-time activity log with color-coded messages
- Meaningful feedback - Connection status, download progress, post-processing steps, errors
- Clean design - Timestamps + colored messages (green=success, orange=warning, red=error)
Smoother Post-Processing Progress
- Animated progress bar - Smooth easing animation instead of jumpy percentage updates
- Better UX - Progress interpolates smoothly between steps (15% easing per frame at 30ms)
v1.0.3
What's New
Real-Time Speed Graph
- Windows 11-style animated speed chart displaying download speed history
- Smooth gradient area fill with forest green theme colors
- Auto-scaling Y-axis adapting to current speeds
- 60-second rolling window of speed data
- Anti-aliased curve with glow effect for crisp rendering
Improved NZB Compatibility
- Obfuscated NZB support: Downloads with scrambled filenames in subjects now clean up properly
- New disk-based fallback detection matches files by release title when NZB parsing fails
- Ensures no leftover RAR/PAR2 files after successful extraction
UI Refinements
- Speed graph panel integrated below progress bar
- Compact Y-axis labels (
80Minstead of80 MB/s) - Reduced padding for tighter, cleaner layout
- Colors aligned with park dark theme (forest green
#217346)
Bug Fixes
- Fixed cleanup skipping for NZBs with obfuscated subject lines
- Files are now correctly identified and removed after extraction
Technical Details
- New
SpeedGraphWidgetclass using pure Tkinter Canvas - New
_derive_filenames_from_disk()method for fallback cleanup - EMA smoothing on speed values for stable graph display
v1.0.2
DLER v1.0.2 - Performance Optimization Release
🚀 Highlights
This release focuses on post-processing performance and user experience improvements, delivering significant I/O optimizations and a cleaner completion workflow.
✨ New Features
FAST PATH Direct Extraction
- Intelligent I/O routing: When download (SSD) and extraction (HDD) directories are on different drives, archives are now extracted directly to the destination — eliminating unnecessary temporary file operations on the SSD.
- Automatic detection: Uses
are_on_same_drive()to determine optimal extraction strategy. - Result: Up to 50% reduction in SSD write operations for cross-drive workflows.
Hardware Auto-Detection
- Automatically detects CPU cores and available RAM to configure optimal parallel extraction settings.
- 64 GB RAM: 4 parallel extractions, 6 threads per 7z process
- 32 GB RAM: 3 parallel extractions, 4 threads per 7z process
- 16 GB RAM: 2 parallel extractions, 4 threads per 7z process
NFO File Handling
- NFO files accompanying archives are now automatically copied to the extracted content if not already present inside the archives.
- Source NFO is deleted from download directory after successful copy to keep the workspace clean.
SFV Cleanup
- SFV checksum files are now automatically cleaned up along with archives and PAR2 files after successful extraction.
🛡️ Safety Improvements
NZB-Specific Cleanup
- Critical safety feature: Cleanup operations now only target files that belong to the current NZB download.
- File names are extracted from the NZB XML structure and used as a whitelist for cleanup.
- If NZB parsing fails: Cleanup is skipped entirely with a warning — manual cleanup required.
- Prevents accidental deletion of files from other downloads or unrelated files in the download directory.
📊 Enhanced Completion Summary
Single unified completion dialog with professional metrics.
- Title adapts: Shows "(Repaired)" only when PAR2 repair actually occurred
- Network metrics: Average download speed in MB/s
- Disk metrics: Extraction throughput between drives
- Critical warnings: Antivirus blocks and cleanup issues shown inline
🔧 Technical Changes
Post-Processor (post_processor.py)
- Added
_is_potentially_nested()for smart extraction path selection - Added
_extract_direct_to_destination()for FAST PATH extraction - Added
_copy_nfo_files()with NZB-aware filtering - Added
_extract_filename_from_subject()for NZB file list parsing - Extended
NZBMetadatawithfilenameslist - Extended
PostProcessResultwith extraction metrics - Added
CLEANUP_SKIPPEDwarning type - Modified
cleanup_archives()to accept whitelist parameter
GUI (tkinter_app.py)
- Added
_show_completion_summary()for unified metrics display - Removed separate warnings popup — integrated into summary
- Pass download metrics to post-processing for complete statistics
Configuration (config.py)
- Version bumped to 1.0.2
📋 Upgrade Notes
- No configuration changes required — new optimizations are automatic
- Existing config files remain compatible
- Hardware detection runs automatically on each post-processing run
🐛 Bug Fixes
- Fixed potential cleanup of unrelated files when multiple downloads share the same directory
- Fixed double I/O on SSD when extracting to a different drive
📦 Download
| File | Size | Description |
|---|---|---|
DLER.exe |
~28.5 MB | Standalone Windows executable (no installation required) |
Requirements: Windows 10/11 x64 with AVX2 support
🙏 Acknowledgments
Thanks to all users who provided feedback on post-processing performance and I/O patterns.
Built with Python 3.14, PyInstaller 6.17, and bundled with par2j64 + 7-Zip for maximum performance.
v1.0.1
First bug squeeze
v1.0.0
High-Performance NZB Downloader for Windows
DLER is a blazing-fast NZB downloader built for modern high-speed connections (1 Gbps+). Featuring a clean, minimalist interface and advanced multithreading capabilities.
Features
Download Engine
- Multi-threaded NNTP with 20+ simultaneous connections
- TLS 1.3 with optimized cipher suites for maximum throughput
- Aggressive pipelining (5-10 requests/connection)
- AVX2-optimized yEnc decoder (~3 GB/s decoding speed)
Post-Processing
- Automatic PAR2 verification and repair
- Smart archive extraction (RAR, ZIP, 7z)
- Password extraction from NZB metadata
- Automatic cleanup after extraction
User Interface
- Modern dark theme (TKinter Modern Themes)
- Real-time speed graph
- Pause/Resume support
- Download queue management
Reliability
- User warnings for incomplete downloads (DMCA/missing articles)
- Antivirus-aware file operations with retry logic
- Post-processing status notifications
What's Included
| File | Description |
|---|---|
DLER.exe |
Standalone executable (no installation required) |
7z.exe + 7z.dll |
Bundled 7-Zip for extraction |
par2j64.exe |
Bundled PAR2 for verification/repair |
System Requirements
- OS: Windows 10/11 (64-bit)
- CPU: x64 with AVX2 support (Intel Haswell+ / AMD Excavator+)
- RAM: 4 GB minimum
- Network: Usenet provider account required
Quick Start
- Download
DLER.exe - Launch and click Settings
- Configure your Usenet server (host, port, SSL, credentials)
- Set download and extraction directories
- Drag & drop NZB files to start downloading
Settings are saved to ~/.dler/config.json
Changelog
v1.0.0 - Initial Release
- Multi-threaded NNTP download engine
- AVX2-accelerated yEnc decoding
- PAR2 verification and repair
- Smart archive extraction with nested RAR support
- Modern TKinter dark theme UI
- Real-time speed graph and ETA
- Pause/Resume functionality
- User warning system for download issues
- Antivirus-resilient file operations
Known Limitations
- Windows only (Linux version planned)
- Requires AVX2-capable CPU
- Single server support (backup servers planned)
License
MIT License - See LICENSE file for details
Made with love for the Usenet community