Why NeatCore?
Intelligent Scan
Streams through folders without freezing the UI. Handles huge trees using batching for smooth responsiveness.
Smart Classification
Identifies images, docs, screenshots, media and archives. Flags low-quality or small images for possible removal.
Duplicate Detection
Finds exact and perceptual image duplicates (pHash) with efficient bucketing and minimal overhead.
Action Recommendations
Suggests delete, move (archive), or compress based on age, location (Downloads), quality metrics and redundancy.
Safety First
Uses Recycle Bin (send2trash) for deletions so you can easily restore mis-clicked files.
Fast Mode
Skips heavy system and build folders (.git, node_modules, AppData) for rapid triage sessions.
Multi-Folder Scan
Select several folders or use Quick Suggest to auto-pick Desktop, Downloads, Documents and more.
Transparent Reasons
Every recommendation includes a clear reason list (age, quality, duplication, folder context).
How It Works
- Select Folders or use Quick Suggest.
- Streaming Scan populates the table as files are seen.
- Analysis classifies each file and computes quality metrics for images.
- Duplicate Pass groups exact + perceptual matches.
- Recommendations applied rules for old screenshots, low-quality images, old downloads, duplicates.
- Bulk Actions choose Delete, Move or Compress safely.
Optional perceptual similarity and AI model loading are deferred to keep startup light.
Get NeatCore
Download the portable folder or run the installer. Built for Windows (64‑bit). Mac/Linux ports can be added later.
Checksum: pending
FAQ
Is it safe?
Deletes go to Recycle Bin. You can restore anytime before emptying it.
Does it use heavy AI?
Core heuristics are lightweight. Optional CLIP model can classify images if you enable it.
Can I undo actions?
Yes—restore from Recycle Bin or keep backups of moved/compressed files.
How are duplicates detected?
Exact hash (MD5) plus perceptual pHash for visually similar images.
Will it auto delete?
No. You explicitly approve actions. Recommendations are guidance only.