STILL: Less Noise, More Stillness | Building a Local-First AI Assistant with Gemma 3

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

What I Built

Meet STILL, a local-first AI assistant that helps you declutter your digital life.

Feed it messy notes, unfinished tasks, or meeting summaries, and let AI organize them into four categories: Act, Keep, Rest, and Archive.

The goal? Spend less time organizing your thoughts and more time on what actually matters.

Demo

Try STILL →

Code

GitHub Repository →

How I Built It

Built with Next.js, TypeScript, Bun, Tailwind CSS, and GSAP, with Gemma 3 running through Ollama for local AI inference.

Gemma 3 analyzes notes, identifies potential priorities, detects overlapping information, and suggests what can wait.

Why Does Open Innovation Matter?

Personal notes can contain sensitive thoughts and unfinished work. Running Gemma 3 locally gives users more control over their data and reduces dependence on paid, proprietary AI APIs.

Open-weight models also let developers experiment with model configurations and build AI tools on their own terms.

The Lazy Payoff

Instead of manually sorting every unfinished thought, let STILL help you find the few things worth your attention.

Not another AI that gives you more work. An AI that helps you do less.

---

Built for Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady.