DEV Community

Aditya Chaudhary
Aditya Chaudhary

Posted on

SlothSync 🦥: Autonomous Anti-Dread File Organizer Powered by Gemma

#devchallenge #hf26challenge #ai #gemma

Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady Submission 🦥

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady

What I Built

In honor of International Sloth Day (October 20), I embraced the sloth philosophy: never do with stressful manual effort what open-weight AI can do quietly in the background.

Every developer, freelancer, and student knows the agonizing dread of digital clutter:

  • The Downloads and Desktop folders turning into chaotic dumping grounds.
  • Cryptically named files like invoice_final_v2(1).pdf, screenshot_2026_10.png, notes_temp_urgent.txt, unfiled code snippets, and duplicate receipts.
  • Sifting, opening, inspecting, renaming, and filing away 50+ files takes 45 to 60 minutes of repetitive, soul-crushing busywork.
  • Because it's tedious, we procrastinate for weeks or months until our workspace is completely unmanageable.

I built SlothSync 🦥 — an autonomous, local-first anti-dread file organizer and backup engine powered by Google's Gemma open-weight models.

How SlothSync Rewards Laziness

Instead of sorting files manually, you activate SlothSync (via our one-click web dashboard or terminal CLI) and relax:

  1. Zero-Effort Semantic Classification: It reads file content snippets and metadata locally using Gemma 2B, automatically categorizing them into clean buckets (Finance, Code, Notes, Media, Archives).
  2. Intelligent Human-Readable Renaming: Converts messy stems (invoice_oct2026_unpaid.txt or copy_of_invoice_final(1).txt) into standardized, searchable names without modifying original contents.
  3. Cryptographic SHA-256 Deduplication: Automatically calculates exact file hashes, flagging identical duplicate files to stop clutter and reclaim wasted storage.
  4. Safety-First Atomic Rollback: SlothSync writes an immutable JSON audit manifest (.slothsync/manifests/) for every session. If you change your mind, a single click on "1-Click Rollback" instantly reverts every file move.
  5. 100% On-Device Privacy: Zero data leaves your machine. Sensitive bank statements, client invoices, and proprietary source code are never uploaded to third-party cloud APIs.

Demo

SlothSync comes equipped with both a sleek, responsive Web Dashboard (zero npm dependencies) and an interactive CLI tool.

1. Interactive Web Dashboard

Run python3 app.py and open http://localhost:8765:

  • Real-Time Sloth Serenity Banner: Tracks active AI model status (Local Gemma 2B vs. Deterministic Heuristics).
  • The Lazy Payoff Widget: Displays live calculated metrics: Minutes Saved, Duplicates Caught, and Dread Level Reduced to 0%.
  • Interactive File Inspector: Displays detected items, categories, tags, and SHA-256 duplicate flags before syncing.
  • 1-Click Rollback: Reverts the entire sync operation atomically in under 1 second.

2. Command Line Interface (CLI) Sample Run 

===========================================================================
  🦥 SlothSync AI - Slow and Steady File Organization
  ---------------------------------------------------
  Rewarding laziness with local-first open-weight AI.
===========================================================================
[*] Scanning source folder: samples/messy_files
[+] AI Engine: Gemma 2B (Ollama Local)
[+] Files Discovered: 4 (0.0 MB)
[+] Duplicates Caught: 1 (SHA-256 matched)
[+] The Lazy Payoff: Saved ~12.8 minutes of manual dread

Files Table:
---------------------------------------------------------------------------
Filename                       | Category     | Suggested Name              
---------------------------------------------------------------------------
copy_of_invoice_final(1).txt   | Finance      | invoice_oct2026.txt [DUPLICATE]
prototype_script_v1.py         | Code         | prototype_script_v1.py      
random_meeting_notes.md        | Notes        | random_meeting_notes.md     
invoice_oct2026_unpaid.txt     | Finance      | invoice_oct2026_unpaid.txt  
---------------------------------------------------------------------------

AI Summary:
"SlothSync handled 4 files (2 Finance, 1 Code, 1 Notes) quietly in the background. Everything is safely organized and backed up without you lifting a finger."
===========================================================================
Enter fullscreen mode Exit fullscreen mode

Code

The complete source code is 100% open-source under the permissive MIT License:

📁 GitHub Repository: https://github.com/AdityaChaudhary5412/SlothSync

Project Architecture & Structure 

flowchart TD
    A["Messy Source Directory (Downloads / Desktop)"] --> B["SlothSync Engine (Web UI / CLI)"]

    subgraph On_Device ["100% On-Device Local Processing"]
        B --> C["SHA-256 Hash & Deduplication"]
        C --> D{"Local Inference Selection"}

        D -->|"Ollama Running"| E["Google Gemma 2B / 9B Open-Weight LLM"]
        D -->|"Offline / Standalone"| F["Deterministic Semantic Classifier"]

        E --> G["Semantic Categorization & Renaming"]
        F --> G

        G --> H["Clean Category Buckets (Finance, Code, Notes, Media, Archives)"]
        H --> I["Atomic Rollback Manifest (.slothsync/manifests)"]
    end

    I --> J["Organized Backup & 1-Click Undo Safety 🦥"]
SlothSync/
├── engine.py              # Local Gemma 2B & heuristic classification engine
├── sync_engine.py         # SHA-256 deduplication, sync & rollback manifest engine
├── app.py                 # Zero-dependency Python web server & REST API
├── cli.py                 # CLI interface for terminal power users
├── static/
│   ├── index.html         # Responsive, dark-mode glassmorphism dashboard
│   ├── style.css          # Theme styles (Forest / Sloth Hacktoberfest palette)
│   └── app.js             # Real-time reactivity, scan & sync handling
├── samples/
│   └── messy_files/       # Sample clutter files for instant evaluation
├── tests/
│   └── test_sync.py       # Automated unit test suite (100% passing)
├── requirements.txt       # Optional extensions (watchdog, rich)
├── LICENSE                # MIT License
└── README.md              # Full documentation
Enter fullscreen mode Exit fullscreen mode

How I Built It

SlothSync was engineered specifically to be lightweight, robust, and zero-barrier:

  1. Local Gemma Inference Engine (engine.py):
    • Connects to local Ollama endpoints (localhost:11434) running gemma2:2b.
    • Low-temperature inference (temperature: 0.2) ensures precise JSON output without conversational noise.
    • Includes a deterministic, pattern-matching heuristic fallback: if someone clones the repo without Ollama running, SlothSync continues to function seamlessly with zero crashes.
  2. Synchronization & Deduplication Engine (sync_engine.py):
    • Employs SHA-256 chunked hashing (64KB buffers) to detect bit-for-bit duplicate files regardless of renaming.
    • Generates structured audit manifests (.slothsync/manifests/manifest_<timestamp>.json) and human-readable Markdown reports (.slothsync/logs/sync_<timestamp>.md).
    • Implements atomic rollback functionality that can revert file operations instantly.
  3. Zero-Dependency Web Server (app.py):
    • Utilizes Python 3's built-in http.server and native REST request handlers.
    • Anyone on macOS, Linux, or Windows can run python3 app.py immediately without running pip install or encountering dependency hell.
  4. Automated Unit Testing (tests/test_sync.py):
    • 100% passing test suite verifying hashing, duplicate detection, categorization heuristics, sync execution, and rollback recovery.

Why Does Open Innovation Matter?

When designing an autonomous tool to manage personal and work files, open innovation and open-weight models are not optional — they are mandatory:

  1. Absolute Data Sovereignty & NDA Compliance:
    • Uploading confidential client contracts, bank records, or source code to proprietary cloud LLMs breaches NDAs and compromises privacy. With open-weight Gemma models, inference runs strictly in local memory.
  2. Zero Marginal Cost ($0 Forever):
    • Cloud AI APIs bill per token. For background daemons watching file system events, API costs quickly accumulate. Gemma runs locally at zero marginal cost.
  3. Immunity to Deprecation & Lock-in:
    • Proprietary models regularly deprecate weights or alter safety boundaries. Open weights are permanent and under the developer's complete control.
  4. Air-Gapped Offline Operation:
    • SlothSync works on airplanes, trains, or off-grid cabins without needing an internet connection.

The Lazy Payoff

Let's quantify the actual laziness payoff of SlothSync:

  • Average time spent manual sorting: ~45 to 60 minutes per month (~10 to 15 hours per year) spent opening cryptic files, creating folders, dragging icons, and second-guessing file names.
  • Time spent with SlothSync: 1.8 seconds.
  • Effort reduction: 96% time saved.
  • Procrastination relief: The mental burden of knowing your workspace is messy disappears completely. SlothSync quietly moves at a "slow and steady" pace, keeping everything organized so you can enjoy your day.

My Agent Session

This project was built, validated, and tested with the assistance of Google Antigravity and DevRelay:

  • Structured prompt engineering for deterministic JSON extraction from Gemma.
  • Architecture audit ensuring zero external runtime dependencies for the core web and CLI modules.
  • Verification of test suites for atomic rollback safety.

Prize Categories

🌟 Best Use of Gemma (Featured Category)

SlothSync is built around Google's Gemma open-weight models (gemma2:2b). Gemma's compact size and high reasoning capabilities make it the ideal local model to run silently on laptops, performing privacy-preserving semantic file analysis and helping developers and creators conquer procrastination.

Happy International Sloth Day and Hacktoberfest 2026! 🦥

Top comments (0)