๐ฆฅ Sloth-OS: The Zero-Effort AI Station that Locks You Out for Working Too Hard
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
What I Built ๐ฆฅ๐ค
I got tired of AI tools demanding high APM, complex prompting, and constant supervision. True laziness isn't just ignoring choresโit's having local AI do them while you rest.
I built Sloth-OS: The Zero-Effort Automation Station. It is a local, AI-powered "Anti-Productivity Suite" designed to automate dreaded daily chores with absolute zero human effort.
Here is what it does:
- ๐จ The Gamified Laziness Monitor: A "Reverse-Tamagotchi" that actively tracks your mouse. If you type too fast or work too hard, it locks your screen with a "CRITICAL: HIGH PRODUCTIVITY DETECTED" alert. You must sit still to earn "Laziness Scores" and unlock it.
- ๐ฌ The Overnight Chat Daemon: Ingests hundreds of missed group chat messages and outputs a strict 2-bullet lazy digest.
- ๐ฅ The Fridge Vision Scanner: Analyzes photos of your sad fridge to automatically generate a missing grocery list and a zero-effort recipe based only on what you have.
- ๐งน The Clutter Vacuum: Automatically categorizes and renames messy desktop files (like
final_FINAL_v2.pdf) into clean folders. - ๐ง Offline Wind-Down Companion: Synthesizes 432Hz brown noise and uses text-to-speech to read AI-generated bedtime stories.
Demo ๐ฅ
Watch the full UI and Local AI in action here:
Code ๐ป
Yash-Ultra / Sloth-OS
An anti-productivity command center built for International Sloth Day and Hacktoberfest. Because true laziness isn't just ignoring choresโit's having local AI do them while you rest.
A neon-soaked fake operating system that quietly does your chores with a local AI model while you nap.
โก Live Desktop Demo
Direct interactive playback: Zero-effort chore engine running live on desktop.
๐ Audio Walkthrough: For narrated audio and explanation, watch the full video on Google Drive โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ SLOTH-OS ยท zero-effort chore engine ยท nap.exe running โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโฌโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค โ overnight_chat_daemon.sh โ fridge_vision_scanner.exe โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโดโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค โ clutter_vacuum.py โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค โ wind_down_synth.wav โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ Contents
- Demo
- What it does
- How it works
- Requirements
- Quick start
- Configuration
- Scripts
- Project layout
- Tests
- Production build
- Privacy and secrets
- Troubleshooting
- License
What it does
Four draggable "windows", each turning a lazy-person chore into one button:
| Window | What you give it | What you get |
|---|---|---|
| overnight_chat_daemon.sh | A chaotic group chat | Two bullet points: what matters, and what to ignore |
| fridge_vision_scanner.exe | A fridge photo (or the sample) | A five-item grocery list and a zero-effort meal idea |
| clutter_vacuum.py |
How I Built It ๐ ๏ธ
Sloth-OS is built using a React + Vite frontend with a heavy emphasis on CSS animations and a glassmorphism UI to mimic a futuristic "Cyberpunk Rainforest" operating system. However, the true core of the project relies entirely on local open-weight models running via Ollama.
- Local Inference Backend: I used Ollama running locally (
http://localhost:8080) as the bridge for all AI tasks. - Text & Sorting (Llama 3): The Chat Summarizer and Clutter Vacuum pass heavy text payloads to the open-weight
llama3model to parse chaos into structured JSON. - Vision (LLaVA): The Fridge Scanner utilizes the open-source multimodal
llavamodel to process uploaded images as Base64 data locally. - Zero-Dependency Audio: The wind-down companion uses the browser's native
AudioContextto synthesize real ambient noise, alongside the nativeSpeechSynthesisAPI to read bedtime stories aloud.
Why Open Innovation Matters ๐
This project relies on the core philosophy of open-source AI: Privacy and Local Execution.
- Total Data Privacy: I am feeding this AI my personal group chats, my messy desktop files, and photos of the inside of my house. Sending that highly sensitive data to a closed-source cloud API is a privacy nightmare. Because this runs locally via Ollama, zero bytes ever leave my machine.
- Costs Nothing to Run: Since I am rewarding laziness, these automation daemons run endlessly in the background. If I used a closed API, I'd be paying per token just to figure out if I need to buy milk. With open-weight models, my ongoing background automation cost is exactly $0.00.
- Total Offline Autonomy: A true lazy companion works from the couch or a cabin in the woods. Because inference is 100% local, I don't even need Wi-Fi on to get my life sorted.
The Lazy Payoff ๐
This week, I used the Overnight Chat Daemon on my college project WhatsApp group. Instead of scrolling through 150 messages of people arguing over deadlines, Sloth-OS gave me: "1. Dave is mad about the server. 2. Push your code by 8 PM." It saved me 15 minutes of scrolling and a ton of mental energy.
Furthermore, the built-in activity monitor successfully forced me to take a 10-minute screen break when my APM got too high while coding this very project. I built an AI that told me to stop working on it.
Prize Categories
- Open-Source AI Core
- Local Agent Innovation