focus-bridge

Sujal Senjaliya — Sun, 04 Oct 2026 19:34:58 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built :-

I built FocusBridge for my friend Alex, an engineering student with ADHD and dyslexia. Reading dense, multi-clause academic papers causes severe working memory fatigue for Alex, making them lose focus and miss core technical mechanics.

FocusBridge is an offline, edge-native reading assistant that turns intimidating papers into digestible mental models:

Visual Ergonomics: Uses client-side Bionic fixation bolding and a guided focus ruler to prevent eye fatigue and visual crowding with zero compute delay.

Local Gemma 2 2B Engine: Deconstructs highlighted text on-device into a one-sentence takeaway, an itemized mechanical breakdown, an analogy anchored in their hobbies, and a quick retention check.

Because it runs completely locally on their laptop, FocusBridge works seamlessly in campus basements and on Wi-Fi-free commutes—delivering instant cognitive support with zero subscription fees and zero data tracking.

Demo

focus-bridge-bice.vercel.app

Code

sujalsenjaliya566 / FocusBridge

How I Built It

FocusBridge is built around Google's Gemma 2 2B Instruct running completely on-device:
1.Inference Runtime: Powered by @mlc-ai/web-llm to execute 4-bit quantized Gemma 2 2B client-side via WebGPU, with an automatic fallback to a local Ollama daemon (gemma2:2b).
2.Frontend Architecture: Built in Next.js 14, pairing a deterministic TypeScript Saccade fixation formatter with a Zustand orchestrator that extracts structured JSON (core takeaways, mechanics, custom analogies, and quizzes) from Gemma with zero external network overhead.

Why Does Open Innovation Matter?

Open innovation is what makes FocusBridge practical for a student:

1.Offline Independence: Alex studies on Wi-Fi-free transit commutes and in campus basement labs where closed APIs simply timeout and break flow state.

2.No Paywall on Accessibility: Cognitive accommodations should not cost $20/month. Gemma’s open weights eliminate recurring subscription fees and per-token API costs.

3.Flow-State Latency & Sovereignty: Running directly on local hardware removes network jitter that shatters ADHD focus, while keeping all reading materials strictly private on Alex's machine.

Prize Categories:-

Featured categories:-($200 for each winner)

Best Use of Gemma -
Use Gemma, Google's open-weight model, in building your project: run it locally, fine-tune it, or serve it through Google Cloud or another provider.

Intro

Sujal Senjaliya — Thu, 01 Oct 2026 13:12:45 +0000

I am new to the community and happy to be part of dev.

DEV Community: Sujal Senjaliya