A hardcore, offline-first productivity OS that turns daily discipline into a gamified tactical loop. Designed for developers and operators, this system uses local Large Language Models (LLMs) to act as an autonomous, private operating system for your habits—guaranteeing zero latency, zero API costs, and 100% data privacy.
Traditional habit trackers rely on cloud APIs and simple checkboxes. Offline Habit Armor treats productivity as a tactical mission. By integrating a local LLM directly into the browser via vanilla JavaScript, the app provides real-time, in-character feedback, dynamic challenge generation, data digestion, and a weighted Gacha reward system—all running completely offline.
- 📝 Dynamic Task Management: Log daily tasks and check them off to earn credits and log deep-work hours.
- 🔥 Boss-Level Challenges: The AI reads your current task list and generates a highly difficult, military-style objective to push your limits.
- 📦 Goods & Provisions Drop(reward system): Spend your earned credits on a weighted RNG loot system to pull for real-world rewards (from an Iced Matcha Latte to Jordan 1 Lows).
- 📊 Interactive Tactical Heatmap: A 28-day visual tracker. Click any cell to log your deep work hours. Watch the matrix-green intensity scale dynamically.
- 📡 AI Tactical Debriefs: The AI analyzes your rolling 7-day work volume and delivers a harsh, analytical performance review.
- 🧠 Master Teaching Lab: Paste dense documentation, codebase snippets, or academic text, and the AI will strictly format it into fundamental mechanics and architectural derivations.
- 🚨 Anomaly Detection: The system silently monitors your 3-day history on load. If you drop your streak, it triggers an urgent background AI alert to break your slump.
- Frontend: Pure HTML5, CSS3 (CSS Grid/Flexbox), and Vanilla JavaScript.
- State Management: Browser
localStorage(No database required). - AI Engine: Ollama (Local HTTP API).
- Default Model:
qwen2.5-coder:7b(Optimized for coding and logical reasoning).
Because this app relies on local AI, you must have an LLM server running on your machine.
Download and install Ollama from ollama.com.
Open your terminal and pull the required model (or your model of choice):
ollama run qwen2.5-coder:7b