This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine
What I Built
trackyAi is a lightweight, intelligent task and reminder tracking application designed to bring order to unstructured daily workflows right from the terminal.
Most task managers feel rigid, demanding strict field inputs, rigid tags, and manual date pickers. trackyAi interprets the theme Ghost in the Machine by acting as an ambient, context-aware digital whisper within your development environment. You feed it raw, conversational thoughts, rough terminal notes, or rapid reminders, and the underlying local intelligence silently parses intent, extracts target deadlines, infers priority, and schedules trackable actions—operating seamlessly in the background without pulling you out of your flow.
Demo
- GitHub Repository: https://github.com/RaqinAlAsraar/trackyAi
- Demo Video / Screenshot:
Code
RaqinAlAsraar
/
trackyAI
Hacktoberfest 26 project
🤖 Tracky AI
Your personal AI assistant for Telegram & WhatsApp.
Manages your daily task lists, timers, calendar events, and proactive reminders naturally through chat and voice.
📌 What is Tracky AI?
Tracky AI understands plain, conversational language just like a human:
- "Create today's task list: prep demo slides, review pull requests, send email"
- "I have an exam on 30 October. Set a reminder for that day at 6:00 am"
- "Change it's timing to 7 am" (instant contextual rescheduling)
- "Show weekly tasklist" (grouped by day)
- "Delete previous tasklists as well which has no use now" (cleans up outdated lists)
Tracky runs 100% locally on lightweight smart NLP (<10ms response time, zero API quota usage), with optional support for local Ollama (
gemma2:2b) or Google Gemini cloud models.
🚀 Quick Setup (3 Minutes)
1. Requirements
- Linux or macOS
- Python 3.10+
- Node.js 18+ (only needed for WhatsApp Web QR linking)
2.
…How I Built It
trackyAi is built entirely with Python and engineered for local-first execution and minimal overhead: - Open-Source AI Core: Utilizes local open-weight models (via Ollama / llama3.2) to perform fast, offline natural language understanding (NLU), entity extraction, and structured JSON intent parsing.
- Architecture: Structured around modular Python handlers for event scheduling, task persistence, and rule-based priority routing.
- CLI & Interface: Built using Rich and custom terminal styling to deliver clear visual dashboards, status indicators, and background system notification hooks.
- Zero Cloud Dependency: Runs completely locally on-device, ensuring zero API latency and absolute data isolation. Why Does Open Innovation Matter? Task lists, personal schedules, and daily dev notes hold sensitive context about your personal life, security workflows, and internal projects. Proprietary cloud-based LLM APIs require you to route private behavioral data to central servers just to parse a simple reminder. Open innovation and open-weight models change the paradigm completely:
- Absolute Privacy: Your personal routines and internal project goals stay strictly on your local machine.
- Offline Resilience: trackyAi works without an internet connection—whether you are working offline, on public transport, or in restricted lab environments.
- Full Customization: Open weights allow fine-tuning and strict system prompt constraints without worrying about sudden API deprecations or usage limits. The Spirit of It There is a distinct feeling of relief when you throw a chaotic mess of thoughts into a terminal prompt at 2 AM after a long debugging session, and an ambient assistant instantly sorts it into an organized, prioritized roadmap. trackyAi turns the cold, clinical command line into an intelligent companion that watches your back, keeps track of open loops, and ensures nothing slips through the cracks. Prize Categories
- Main Track: Open-Source AI Challenge — Week 4: Ghost in the Machine
- Local AI / Open-Weight Models: Best application built on locally executed, open-weight language models.
Top comments (0)