DEV Community: Aishwarya

Sloth Digest: Slow news. Taken at your own pace.

Aishwarya — Sun, 25 Oct 2026 12:17:42 +0000

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

What I Built

Demo

accounts.google.com

Code

aishwarya150624 / sloth-digest

app to help lazy and boring tasks

🦥 Sloth Digest

Wake up to a 2-line summary of your group chats, written by an AI that never leaves your laptop.

Export a chat (WhatsApp: Chat → More → Export chat → Without media), drop the .txt in a folder, and run one command. Sloth Digest reads only the new messages since last time and writes a digest:

## family_group (13 new messages)
TL;DR: <two short sentences>
Plans
- <dates, places>
Needs you
- [ ] <things asked of you>

Why open-source AI?

Private: your family chats never touch a cloud API. Summarizing them with a hosted model means handing a stranger your conversations.
Free forever: a nightly run over many chats would add up on a paid API. Locally it costs electricity.
Offline: works without internet.
Swappable: --model qwen2.5:7b for sharper summaries, or a small model for an old laptop.

Built for laziness

Incremental: remembers what it…

View on GitHub

How I Built It

🛠️ How I Built It

Building Sloth Digest started with a simple idea: What if technology helped us slow down instead of constantly demanding our attention?

For the Hacktoberfest Week 3: Slow and Steady challenge, I wanted to create something that made information consumption simpler, calmer, and more intentional.

🦥 1. Designing a Slower Experience

I began by defining the core philosophy of Sloth Digest: quality over quantity.

Instead of creating another traditional news feed, I designed the experience around a single weekly issue containing meaningful stories, interesting insights, and positive developments.

The goal was to eliminate unnecessary browsing and make staying informed feel effortless.

📰 2. Structuring the Weekly Digest

Each issue follows a simple reading flow:

Weekly Snapshot → Curated Stories → Reading Progress

The weekly snapshot provides a quick overview, while curated articles encourage readers to explore topics more deeply.

A gentle progress

Why Does Open Innovation Matter?

🌍 Why Does Open Innovation Matter?

Because meaningful innovation should be accessible to everyone, not locked behind expensive subscriptions, proprietary systems, or closed ecosystems.

While building Sloth Digest, I wanted to explore a different relationship between technology and attention. Instead of creating another platform designed to keep people scrolling, I envisioned something that helps them consume less information but understand more.

That's where the philosophy of open innovation becomes especially meaningful.

🦥 Freedom to Build Differently

Open-source technology gives developers the freedom to experiment, customize, and improve their solutions without being entirely dependent on a single company's infrastructure.

For a project like Sloth Digest, this flexibility matters.

AI Can Write Your Proposal. But Can It Prove What It Says? Meet PITCHCRAFT.

Aishwarya — Sun, 18 Oct 2026 03:27:42 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

Demo

accounts.google.com

Code

aishwarya150624 / Pitchcraft

PITCHCRAFT — An evidence-first AI proposal studio for freelancers. Create personalized, portfolio-backed proposals, verify claims, and track opportunities.

PITCHCRAFT

A local-first proposal workspace for freelancers. Paste a client brief and get a proposal drafted only from your own profile and portfolio. Every sentence is traced to its source, and anything unbacked is flagged. No build step, no dependencies, no backend. Data lives in your browser.

Run it

Open this folder in VS Code and accept the recommended extensions.
Right-click index.html and choose Open with Live Server (port 5500), or run npm start (port 3000).
Press F5 to open it in Chrome with the included launch configs.

Opening index.html directly from disk also works. The Google Fonts link needs an internet connection without it the page falls back to Georgia and system fonts.

Structure

File	Purpose
`index.html`	Page shell and script order
`css/styles.css`	All styling, with light and dark theme tokens
`js/state.js`	Sample data, localStorage persistence, small helpers
`js/engine.js`	`check` (claim verification), `analyse` (brief to needs, fit, draft), `segs`

…

View on GitHub

How I Built It

Why Does Open Innovation Matter?

Prize Categories

🌿 WANDER — An AI App That Wants You to Go Outside An app that gives you a reason to go outside, then gets out of the way.

Aishwarya — Sun, 11 Oct 2026 05:49:59 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

🌿 What I Built

What if an AI app encouraged you to stop using it?

Meet WANDER — an offline-first AI companion that turns everyday surroundings into little adventures.

Powered by Google's Gemma 3 4B through Ollama, WANDER uses local AI to generate three personalized outdoor missions based on your mood, available time, weather, and previous explorations. Gemma helps keep each adventure fresh and relevant without relying on cloud-based AI.

🌱 Choose your mood and available time.

✨ Discover three unique missions generated by Gemma 3.

📵 Disconnect as your screen goes dark.

🌍 Explore your surroundings and rediscover overlooked details.

📖 Reflect in your private Wander Journal and track your Wander Ratio — time outside versus time on screen.

WANDER doesn't measure success by how long you stay, but by how much you experience after leaving. 🌿

Demo

accounts.google.com

Code

aishwarya150624 / WANDER

🌿 WANDER — An offline-first AI companion powered by Gemma 3 that generates personalized outdoor missions to help you disconnect from screens and rediscover the real world.

WANDER (Day 1 build)

Run

ollama pull gemma3:4b (any model name works: WANDER_MODEL=qwen2.5:7b)
pip install -r requirements.txt
Optional weather: export WANDER_LAT=.. WANDER_LON=..
uvicorn main:app --host 0.0.0.0 --port 8010
On your phone (same Wi-Fi) open http://<laptop-ip>:8010

What exists

/api/tasks: 3 tasks, schema-enforced, random ingredient, history-aware, safety check, repeat filter
/api/finish, /api/journal: SQLite journal with wander ratio
safety_log table keeps every rejected task, for the "what the checker caught" count

Known gaps (Day 2)

Photo + caption, season layer, request log / privacy receipt, eval scripts
Screen time counts from "Get three tasks" to saving, not the setup screen before it
Not yet run against a real model: the first real run will show if prompts need tuning

Day 2 additions

Photo on the return screen: resized on the phone, captioned by the local vision model, never judged
Season layer: prompt hint by date (autumn: leaf colour stages…

View on GitHub

How I Built It

🛠️ How I Built It

I designed WANDER around a simple idea: AI should encourage exploration, not endless interaction.

The project uses an offline-first architecture with the following technologies:

*Gemma 3 4B * Runs AI locally to generate personalized outdoor missions based on mood, time, weather, and previous activities.
SQLite: Stores mission history and private journal entries locally.
HTML, CSS & JavaScript: Power the lightweight, mobile-friendly interface

The AI generates three missions, checks them for safety and repetition, and lets users choose one before putting their phones away.

I also designed a Wander Ratio to measure time spent outdoors versus time on screen, alongside a Privacy Receipt feature to make data usage transparent.

The biggest focus was keeping WANDER private, lightweight, and genuinely useful beyond the screen. 🌿

Why Does Open Innovation Matter?

🌍 Why Does Open Innovation Matter?

I believe technology should give us more freedom, not demand more of our attention.

Open innovation is what makes WANDER possible. By using Gemma 3 through Ollama, WANDER can generate personalized outdoor missions locally without depending on expensive cloud APIs.

Here's why it matters:

🔒 Privacy First: Personal reflections, photos, and activity history can stay on the user's device.
🌐 Offline Freedom: Core AI features can work without an internet connection.
🔄 Flexibility: Developers can experiment with different open-weight models and customize the experience.
💚 Accessibility: No per-request API fees, making the project easier to explore and improve.

A closed API wouldn't offer the same level of offline independence, privacy, and control.

For me, open innovation isn't just about sharing code. It's about building AI that respects people's privacy, choices, and time. 🌿

Prize Categories

Best Use of Gemma

WANDER uses Google's Gemma 3 4B open-weight model, running locally through Ollama, to generate personalized outdoor missions based on users' moods, available time, surroundings, and previous activities.

Gemma is central to WANDER's goal of making AI-powered exploration private, offline-first, and accessible without relying on paid cloud APIs.

STAGE LADDER - A Private, Offline Speaking Coach Built for a Friend

Aishwarya — Mon, 05 Oct 2026 01:51:30 +0000

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

What I Built

Demo

Code

How I Built It

Why Does Open Innovation Matter?

Prize Categories

STAGE LADDER - A Private, Offline Speaking Coach Built for a Friend

Aishwarya — Sun, 04 Oct 2026 16:20:28 +0000

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