DEV Community: Aditya Kumar

🧠 Echoe — A Second Brain Simulator

Aditya Kumar — Sat, 31 Oct 2026 15:18:14 +0000

Ghost in the Machine 👻

What if your memories didn't just disappear into folders, forgotten notes, and old documents?

What if they could think, connect, and resurface when you need them?

Echoes is an AI-powered Second Brain Simulator built around Gemma 4. The idea is to create a searchable, interconnected memory system that doesn't simply store information but understands relationships between it.

👻 What does it do?

Memory Ingestion: Collects notes, PDFs, conversations, research, and other personal information.
Knowledge Graph: Connects people, concepts, events, facts, and ideas.
Associative Recall: Retrieves memories through meaning and connections rather than just keywords.
Memory Timeline: Revisits information based on when it happened.
Contradiction Detection: Identifies conflicting information across stored memories.
Dream Mode: An experimental feature that explores forgotten connections and discovers unexpected relationships between memories.

⚡ Why Open Source AI?

The idea behind Echoes is that your memories should belong to you.

By using an open-weight model like Gemma 4, the project aims to support local inference, private data processing, model customization, and independence from proprietary AI APIs.

Your personal knowledge shouldn't have to live on someone else's server.

Open models also make it possible to experiment with how AI processes, retrieves, and connects information without being locked into a single provider.

🛠️ Tech Direction

AI Model: Gemma 4
Foundation: Sproutpatch (fork-based development)
Core Concept: Local AI + Knowledge Graph + Associative Memory
Goal: A personal memory system accessible across laptop and phone.

💭 The Bigger Idea

We spend years collecting information but very little time reconnecting with it.

Echoes explores the possibility of an AI that doesn't just answer questions but helps you rediscover things you've forgotten.

A second brain that never truly forgets.

Because sometimes, the most interesting ghost in the machine is a memory you thought was gone.

Hacktoberfest #OpenSourceAI #Gemma #GhostInTheMachine #LocalAI #ArtificialIntelligence

TouchGrass AI: Turning Screen Breaks Into an AI-Powered Side Hustle

Aditya Kumar — Wed, 14 Oct 2026 16:50:02 +0000

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

What I Built

TouchGrass AI is an open-source outdoor activity generator that helps students and remote workers take meaningful breaks from their screens.

Users can generate personalized outdoor missions based on their available time and preferences, complete challenges, and track their progress.

The goal is to turn a simple habit into a useful product that people would actually want to use.

How I Built It

I built TouchGrass AI using HTML, CSS, JavaScript, Python, and FastAPI.

The AI component uses an open-weight language model through Ollama to generate personalized outdoor activities. SQLite stores mission history and completion records.

[Adjust this section to match the implementation you actually completed.]

Why Does Open Innovation Matter?

Open-source AI makes TouchGrass AI more accessible and gives developers control over how the product works.

Running an open-weight model locally can reduce dependence on paid AI APIs and keep personal activity data on the user's machine. It also allows the model to be replaced or customized as the project evolves.

For a small side project, this makes experimentation easier and gives the product a path toward sustainable operating costs.

TouchGrass AI: An Open-Source AI That Gets You Outside 🌿

Aditya Kumar — Fri, 09 Oct 2026 18:50:55 +0000

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

TouchGrass AI 🌿: An Open-Source AI That Gets You Outside

What if AI could help you spend less time staring at a screen and more time exploring the world around you?

That's the idea behind TouchGrass AI — a simple outdoor activity generator powered by open-source AI. Instead of recommending another app to scroll through, it gives you a reason to put your phone away and go outside.

What I Built

TouchGrass AI turns everyday outdoor time into small, achievable adventures.

Users can choose how much time they have available and receive a personalized outdoor mission, such as:

10 minutes: Find three different leaf shapes around you.
20 minutes: Take a walk and observe five things you usually overlook.
30 minutes: Explore a nearby green space and record something interesting you discover.

The goal isn't to spend more time using an app. It's to make the app useful enough that you can close it and enjoy the real world.

TouchGrass AI is designed for students, people who spend long hours at their computers, and anyone looking for a simple reason to step outside.

Demo

🚀 Live Demo: [Add your deployed application URL]

🎥 Video Demo: [Add your demo video URL, if available]

The demo will show how a user generates a mission, completes the outdoor activity, and records their experience.

Code

💻 GitHub Repository: [Add your GitHub repository URL]

The project source code, setup instructions, and AI configuration will be available in the repository.

How I Built It

I built TouchGrass AI around the idea that AI should encourage real-world experiences rather than demand more screen time.

The planned technology stack includes:

Python and FastAPI: Backend and AI integration.
Ollama: Running a compatible open-weight language model locally.
HTML, CSS, and JavaScript: The user interface and interactive experience.
SQLite: Storing completed missions and activity history.

The AI generates outdoor missions based on the user's available time, surroundings, and preferences. The application then presents a simple task that the user can take outside.

The focus is on keeping the experience lightweight: generate a mission, leave the screen, complete the activity, and return only if you want to record it.

Implementation note: Update this section to match the technologies and features actually working in your submitted version.

Why Does Open Innovation Matter?

For this project, open innovation is about more than choosing a different AI model.

Using an open-weight model that can run locally offers several advantages:

Privacy: User preferences and prompts can be processed locally instead of being sent to a third-party AI service.

Accessibility: Running a compatible model locally can eliminate per-request API charges, although hardware and electricity still have costs.

Freedom to experiment: Different models and prompts can be tested to find the best balance between creativity, speed, and useful recommendations.

Transparency and control: Developers can inspect the surrounding application logic, customize how missions are generated, and adapt the experience to their needs.

A closed API could generate outdoor activities too. However, a locally running open-weight model gives this project more control over how AI is used and where the data goes.

Open innovation makes it possible to build an AI experience that is customizable and potentially usable without an internet connection.

My Agent Session

[Optional: Add your DevRelay agent session link here.]

This section can document how the project was developed, including the prompts, decisions, experiments, and debugging steps involved.

Prize Categories

[Add the applicable partner categories you are entering, based on the official challenge page.]

Final Thoughts

Technology doesn't always need to keep us connected to a screen.

Sometimes, its best use is helping us discover something beyond it.

TouchGrass AI is my attempt to explore that idea with open-source AI: a small tool designed to turn a few minutes of free time into an opportunity to move, observe, and reconnect with the world around us.

Now it's time to close the laptop and touch some grass. 🌱

Thanks for reading!

A BEGINNING....

Aditya Kumar — Mon, 05 Oct 2026 11:52:14 +0000

Excited to share my work