DEV Community: Anwesha

DraftLight: A little light for your unfinished drafts

Anwesha — Sat, 31 Oct 2026 04:49:21 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine

What I Built

A little light for the things you haven't finished.

We all have unfinished things.

A story we stopped writing. A project we abandoned halfway. An idea we were once excited about but somehow lost track of. Sometimes, the hardest part isn't doing the work — it's finding our way back to it.

That's the idea behind DraftLight.

DraftLight is a local-first AI companion for returning to unfinished work. It helps you understand where you left off, recognize what's unfinished or unclear, and find a practical next step.

It's not just another writing assistant or productivity tool. It's a space for work that has lost its momentum.

Whether you're working on a story, a software project, a personal goal, a plan, or a creative idea, DraftLight helps you reconnect with the context you left behind.

The goal is simple: make returning feel a little less overwhelming.

Demo

A live demo will be available soon.

I'll add the deployment link here once it's ready!

Code

Explore DraftLight on GitHub:

anweshasingh-dev/DraftLight

How I Built It

DraftLight is designed around the idea of a local-first AI companion.

Rather than treating unfinished work as something that needs to be rewritten or replaced, the AI helps you make sense of what already exists.

The core experience revolves around a simple text workspace where you can preserve the context of your work and use AI to help you:

Understand where you left off.
Identify unfinished, unclear, or disconnected parts.
Recover the direction of an idea.
Find a manageable next step.

The project explores how AI can become a thoughtful companion to our work, rather than simply generating more content for us.

Why Does Open Innovation Matter?

I believe unfinished work is deeply personal.

Our drafts, ideas, projects, and plans often contain pieces of our lives that we may not want to hand over to a third-party service.

That's why local-first AI matters to DraftLight.

Open innovation makes it possible to explore AI experiences that offer greater control over personal context, encourage transparency, and give developers the freedom to build tools around people's actual needs.

It also makes the technology more approachable. Developers can experiment, adapt, and improve the experience without being entirely dependent on a closed API.

For DraftLight, the vision is to make AI feel less like an external service and more like a companion that helps you reconnect with your own work.

The Spirit of It

There's something quietly heartbreaking about opening a project you once cared about and realizing you don't remember where you were going with it.

Maybe you stopped because life got busy. Maybe you lost motivation. Maybe you simply didn't know what to do next.

But unfinished doesn't necessarily mean forgotten.

Sometimes, all you need is a little help finding your way back.

That's what DraftLight is meant to be.

Not another tool telling you to be more productive. Not another reminder of everything you haven't done.

Just a little light for the things you haven't finished.

PencilPear 🍐✏️ a tiny shared space for friends to leave a little piece of themselves.

Anwesha — Sun, 25 Oct 2026 07:45:37 +0000

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

What I Built

Have you ever wanted to share a random thought, doodle something silly, or leave a little note for a friend without starting another conversation?
A shared digital sticky-note workspace for forgetful study buddies. Drop reminders, sketch ideas, and yell "DO THE ASSIGNMENT" right onto each other's desktop screens in real time.
That's the idea behind PencilPear.

I built PencilPear for a friend as a lightweight collaborative sticky-note app where we can write, draw, and erase on the same shared board in real time. Instead of sending every little thought as a separate message, we can simply open our board and leave something there.

Each room has its own board, so different spaces stay separate. The board state persists, allowing us to return to what we've created, and the app integrates with Electron for a desktop experience.

It's a small project, but the idea behind it is simple: friendship doesn't always need a big conversation. Sometimes, a tiny doodle is enough.

Demo

Try PencilPear here: pencilpear.vercel.app

Code

Explore the source code on GitHub:

anweshasingh-dev/pencilpear

How I Built It

I built PencilPear with help from OpenAI Codex as my AI coding assistant.

The application brings together a shared drawing and writing board, real-time collaboration, room-based separation, persistent board state, and Electron integration for a desktop app experience.

I worked toward keeping the experience lightweight and focused on one thing: making it easy for two friends to share a small creative space without unnecessary complexity.

Codex helped me work through the implementation, but the product idea came from a simple question: What if staying connected with a friend could feel as easy as leaving a sticky note on their desk?
note: PencilPear does not uses an open-weight AI model or an open-source AI framework inside the application.

Why Does Open Innovation Matter?

Open innovation makes it easier for developers to experiment, learn, and turn small ideas into working projects without needing a large team or a big budget.

For a project like PencilPear, that matters because the goal isn't to build another complicated communication platform. It's to explore a small, personal idea and make it real.

Using an AI coding assistant helped me move from an idea to a publicly accessible project. Having the source code available on GitHub also means other developers can explore the implementation, learn from it, and potentially build on it.

PencilPear is a reminder that useful software doesn't always have to solve a massive problem. Sometimes, it just needs to make one little thing better for someone you care about.

NeatNest: A Little Less Mess, A Lot More Organized 🪺

Anwesha — Sun, 25 Oct 2026 07:36:28 +0000

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

What I Built

Have you ever downloaded a file, saved it somewhere, and then completely forgotten where it went?

Screenshots, PDFs, documents, images, installers, and random downloads can quickly turn a computer's folders into a mess. Organizing everything manually takes time, and putting files in the wrong place can make them even harder to find later.

That's why I built NeatNest — a lightweight desktop file organizer designed to make managing files simpler.

NeatNest helps users identify files that need organizing, review suggested destinations, and bring a little order to their cluttered folders without having to sort everything from scratch.

The idea is simple: spend less time managing files and more time doing the things you actually downloaded them for.

Demo

🎥 Demo: https://neat-nest-liard.vercel.app/

Code

🔗 GitHub Repository: https://github.com/anweshasingh-dev/NeatNest

NeatNest is being developed with a focus on simplicity, usability, and keeping everyday file management straightforward.

How I Built It

I built NeatNest as a desktop utility focused on making file organization less tedious.

Rather than immediately moving files without the user's involvement, the workflow focuses on suggesting how files could be organized and letting users review those suggestions before taking action.

The project focuses on three things:

File discovery: Identify files that could benefit from better organization.
Organization suggestions: Recommend suitable destinations based on file information.
User-controlled organization: Help users act on those suggestions instead of manually figuring out where every file belongs.

I used AI-assisted development tools to help build and refine the application, breaking implementation into smaller tasks so each part could be tested and improved without making the project unnecessarily complicated.

For the AI component, NeatNest's current organization workflow should be described according to the actual implementation. If it uses an open-weight model or an open-source AI framework, name it here and explain exactly what it does.

Why Does Open Innovation Matter?

File organization is an everyday problem, and solving it shouldn't require an unnecessarily complicated or expensive tool.

Open innovation makes it easier to experiment, learn from existing tools, inspect implementation choices, and build utilities that people can adapt to their own workflows.

For a project like NeatNest, transparency is especially important because the application works with personal files. Users should be able to understand how their files are handled and have control over what gets organized.

My goal is to keep building something practical, approachable, and useful — one improvement at a time.

The Lazy Payoff 💤

Imagine downloading a bunch of files and not having to sort through every single one manually.

Instead of repeatedly opening folders, checking file types, and deciding where everything belongs, NeatNest helps streamline the process by suggesting an organization structure.

The payoff isn't about automating your entire computer. It's about removing a small, repetitive chore that adds up over time.

Because sometimes, productivity means doing less unnecessary work.

Prize Categories

[Add the applicable partner prize categories here, based on the challenge rules.]

Trajectory: A More Realistic Way to Plan Your Goals

Anwesha — Sun, 18 Oct 2026 05:46:29 +0000

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

What I Built

Trajectory is a work-in-progress project exploring a simple question: What if planning your day started with your goals, not just a list of tasks?

Students juggle courses, personal projects, assignments, and hobbies. Independent learners try to build new skills while managing existing responsibilities. It's easy to create an ambitious to-do list, but much harder to decide what deserves attention first and how much time to give it.

Trajectory is my attempt to make that process more realistic.

The initial version focuses on two core features:

Goal-to-Plan: Describe a bigger goal and what you want to achieve. AI helps turn it into a structured plan with actionable steps and milestones.
Daily To-do Planner: List the tasks you want to work on, assign each a deadline (2 days, 4 days, 1 week, or no deadline), and provide your available time. AI suggests a priority order, reasonable time allocations, and why each task matters.

The idea is to keep tasks and progress available across days while generating a fresh daily plan based on current priorities and unfinished work. Rather than treating every day as a clean slate, Trajectory aims to help users make steady progress toward their goals.

The longer-term vision includes more adaptive planning, progress-based forecasting, personalized career guidance, and a creative-practice journal where users can save sketches and reflect on their artistic development. These are future possibilities, not claims about the current implementation.

Demo

Project demo: https://trajectory-git-main-mystic-force.vercel.app/

The project is currently in development. The demo will be updated as the core features become functional and ready to showcase.

Code

GitHub repository: https://github.com/anweshasingh-dev/trajectory

How I Built It

The planned technology stack includes:

Python and Flask for the backend.
SQLite for persistent storage of goals, tasks, and progress.
HTML, CSS, and JavaScript for the user interface.
An open-weight instruction model, with a suitable inference setup to be selected and verified, for interpreting natural-language goals and proposing structured plans.
Python application logic to validate inputs, enforce planning constraints, and manage task progress.

An important design principle is to keep AI suggestions separate from application-controlled logic. The model can help interpret goals, break work into actionable steps, and explain recommended priorities. The application should remain responsible for validating data and preserving user-approved changes.

The initial focus is deliberately small: generate a goal-based plan, prioritize daily to-dos, and retain progress between sessions. More advanced scheduling and forecasting can be explored once this foundation works reliably.

The final implementation, model choice, and verification results will be documented as development progresses.

Why Does Open Innovation Matter?

Personal planning involves information that can be sensitive, including goals, schedules, unfinished work, and learning habits.

Open-weight models and local inference offer an opportunity to experiment with AI-assisted planning without making every interaction dependent on a closed, externally hosted API.

For Trajectory, this approach could support:

More control: Developers can experiment with compatible models and adapt the system to their needs.
Privacy-conscious design: Local inference can keep prompts on the user's machine, provided the application and its other data flows are designed accordingly.
Accessible experimentation: Students and independent developers can explore AI-powered products without making a paid proprietary API the foundation of every interaction.

I want the AI to serve a clear purpose rather than simply add a chatbot to another to-do list. The aim is to explore how open models can help turn goals into actionable work while transparent application logic keeps the planning process grounded and manageable.

The open-weight model and inference setup are still being evaluated, so their final implementation and limitations will be reported honestly.

What's Next?

The immediate goal is to complete and test the two core features: generating goal-based plans and creating prioritized daily to-do plans with persistent progress.

From there, I want to learn what works in real planning scenarios before expanding the project. Trajectory is a small first step toward a more personal planning system that connects everyday actions with longer-term ambitions.

🌱 Little Elsewhere - A Little More Adventure, A Little Less Screen Time

Anwesha — Sun, 11 Oct 2026 04:50:20 +0000

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

What I Built

Little Elsewhere 🌱

I have little cousins who are glued to their phones most of the time, and it got me thinking, what if going outside felt as exciting as playing a game on your phone?

So I started building Little Elsewhere, a small adventure game that nudges kids to put their screens down and explore the real world instead.

Kids can choose their explorer avatar, pick a mood, try nature bingo, complete little outdoor quests, and collect rewards for their adventures.

The idea is simple: instead of just telling kids to stop using their phones, give them something fun to look forward to outside.

It's still a work in progress, but that's the direction I'm building toward. 🌿

Demo

🔗 Live app:

(https://little-elsewhere-one.vercel.app/)

Code

💻 GitHub repository: (https://github.com/anweshasingh-dev/little-elsewhere)

How I Built It

I built Little Elsewhere using React, TypeScript, and Vite.

I used Codex as a coding assistant while working through the project in small steps, refining the UI, organizing the code, and testing the features as I went.

Some of the current features include:

Explorer avatars and mood selection
Nature bingo
Predefined outdoor quests
Coins and rewards
Local persistence for saved progress

I'm still improving the experience and figuring out how to make it more engaging for kids without making the interface cluttered.

Why Does Open Innovation Matter?

Kids don't necessarily need another app asking them to spend more time on a screen. Sometimes, they just need a fun reason to step outside.

I'm building this as an open-source project because I'd love for other people to experiment with the idea and make it better.

Open innovation also means other developers can contribute new quests, improve accessibility, suggest better ways to engage kids, and help the project grow beyond what I can build alone.

What's Next?

Little Elsewhere is still evolving. I'd like to expand the quest collection, improve the reward system, and explore ways to make outdoor activities more personal and accessible.

For now, this is my first step toward turning a small observation about my cousins into something other families might enjoy too. 🌱

Thanks for reading!