local-file-organizer 2.0.0a2

AI-powered local file management with state-of-the-art models

File Organizer v2.0

AI-powered local file management. Privacy-first -- runs 100% on your device.

307 tests | 334 modules | 48+ file types | Python 3.11+

Features

• AI-Powered Organization: Qwen 2.5 3B (text) + Qwen 2.5-VL 7B (vision) via Ollama
• Audio Transcription: Local speech-to-text with faster-whisper (GPU-accelerated)
• Video Analysis: Scene detection and keyframe extraction
• Copilot Chat: Natural-language assistant -- "organize ./Downloads", "find report.pdf", "undo"
• Organization Rules: Automated sorting with conditions, preview, and YAML persistence
• Terminal UI: 8-view Textual TUI (Files, Analytics, Audio, History, Copilot, and more)
• Web UI: Browser-based interface via FastAPI and HTMX
• Full CLI: Organize, rules, suggest, dedupe, daemon, analytics, update, profiles
• Auto-Update: GitHub Releases checks with verified downloads and rollback
• Intelligence: Pattern learning, preference tracking, smart suggestions, auto-tagging
• Deduplication: Hash and semantic duplicate detection
• Undo/Redo: Full operation history
• PARA + Johnny Decimal: Built-in organizational methodologies
• Cross-Platform: macOS (DMG), Windows (installer), Linux (AppImage) executables

Screenshots

Quick Start

pip install -e .

# Pull models
ollama pull qwen2.5:3b-instruct-q4_K_M
ollama pull qwen2.5vl:7b-q4_K_M

# Organize files (dry run first)
file-organizer organize ./Downloads ./Organized --dry-run

# Launch the TUI
file-organizer tui

Web UI (Preview)

Start the FastAPI server and open the UI:

uvicorn file_organizer.api.main:app --reload

Then visit http://localhost:8000/ui/ for the HTMX interface.

Documentation

• User Guide
• CLI Reference
• Configuration Guide
• Troubleshooting
• Getting Started

Optional Feature Packs

Pack	Install Command	Features
Audio	pip install -e ".[audio]"	Speech-to-text (faster-whisper, torch)
Video	pip install -e ".[video]"	Scene detection (OpenCV, scenedetect)
Dedup	pip install -e ".[dedup]"	Image deduplication (perceptual hashing)
Archive	pip install -e ".[archive]"	7z and RAR archive support
Scientific	pip install -e ".[scientific]"	HDF5, NetCDF, MATLAB formats
CAD	pip install -e ".[cad]"	DXF and CAD format support
Build	pip install -e ".[build]"	Executable packaging (PyInstaller)
All	pip install -e ".[all]"	Everything above

Audio system dependencies

For full audio format support, the [audio] pack uses FFmpeg (all platforms) and optionally CUDA + cuDNN (NVIDIA GPU users).

FFmpeg — required for non-.wav formats (MP3, M4A, FLAC, OGG); optional if you only transcribe raw .wav:

# macOS
brew install ffmpeg

# Ubuntu / Debian
sudo apt install ffmpeg

# Windows (winget)
winget install ffmpeg

CUDA + cuDNN — optional, for significantly faster transcription (see faster-whisper benchmarks for hardware-specific numbers):

# Install CUDA Toolkit from https://developer.nvidia.com/cuda-downloads
# Install cuDNN from https://developer.nvidia.com/cudnn

# Verify the full transcription backend (not just PyTorch)
python3 -c "from faster_whisper import WhisperModel; print('faster-whisper OK')"
python3 -c "import torch; print('CUDA:', torch.cuda.is_available())"

Fallback behavior: without FFmpeg, only .wav files are transcribed; other formats are organized by filename/metadata but not content-analyzed. Without CUDA, transcription runs on CPU (slower but fully functional).

See the Installation Guide for troubleshooting and advanced configuration.

Development

# Run tests
pytest

# Lint
ruff check src/

Configuration

Config lives in config/file-organizer/config.yaml relative to your config home. Override with FILE_ORGANIZER_CONFIG.

---

Status: Alpha | Version: 2.0.0-alpha.1 | Last Updated: 2026-03-01