DEV Community: Nidhi Jha

ECHOEVER: Every Memory Deserves an Eternity | Building a Digital Memory Archive with Open-Source AI

Nidhi Jha — Sat, 31 Oct 2026 13:14:14 +0000

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

What I Built

What if the stories, photographs, and little moments we never want to forget could live somewhere forever?

That's the idea behind ECHOEVER, an AI-powered digital memory preservation platform built for the Ghost in the Machine challenge.

ECHOEVER is designed to help people preserve their meaningful experiences, photographs, personal stories, and memories in one interactive digital archive.

Instead of letting precious memories get lost in scattered folders, old photographs, or forgotten notes, ECHOEVER brings them together in a space where they can be explored, organized, and rediscovered.

The project features:

🕯️ Digital Memory Archive: Create, organize, and preserve personal memories.

✍️ Memory Creation Studio: Transform simple notes into beautifully written stories with AI assistance.

👻 AI Memory Companion: Ask questions about saved memories and rediscover meaningful moments.

⏳ Interactive Timeline: Explore memories through a chronological visual experience.

🔐 Privacy-First Approach: Keep personal memories under your control through a local-first architecture.

Why Ghost in the Machine?

Halloween and Día de los Muertos remind us of something meaningful: even when moments pass, the stories and memories associated with them remain.

I wanted to interpret the theme differently. Instead of creating a chatbot that pretends to communicate with ghosts, I wanted to build something that gives memories a digital presence.

ECHOEVER is about preserving the echoes of the past without pretending to bring the past back.

Demo

Code

Nidhijha-1001 / memento

ECHOEVER

Every Memory Deserves an Eternity.

A private, local-first home for the moments you choose to keep.

Photographs fade. Stories get scattered. ECHOEVER gives them a place to return to.

Branding note: ECHOEVER is the project name used in this repository documentation. The current application UI and API still display their working name, Echoes Beyond; this README-only update does not rename the website.

The idea

ECHOEVER is an AI-assisted digital memory archive for photographs, personal stories, and meaningful experiences. It helps you record a moment, keep related memories together, rediscover details through search or a timeline, and—when you choose—ask a local open-weight model to help shape your notes into…

View on GitHub

How I Built It

I built ECHOEVER using a full-stack architecture.

Frontend

React.js

Vite

TypeScript

Tailwind CSS

Framer Motion

Backend

Python

FastAPI

SQLAlchemy

SQLite

Open-Source AI

Ollama for local model inference.

Gemma / Qwen open-weight models (depending on the configured model).

The architecture separates the frontend, backend, database, and AI service.

The frontend communicates with FastAPI, which handles memory management and AI requests. The AI service communicates with Ollama to run inference locally.

The goal is to make AI functionality available without depending on a proprietary hosted AI API.

Why Does Open Innovation Matter?

For ECHOEVER, open innovation is not just a technical preference. It is connected to the central purpose of the project.

Personal memories are deeply private. Photographs, family stories, and personal experiences should not have to be sent to an external AI provider simply to make them searchable or meaningful.

Using open-weight models and local inference offers several advantages:

Privacy and ownership

Local inference makes it possible to process personal memories without sending prompts and memory context to a third-party hosted AI service.

Freedom to experiment

Open-weight models make it possible to experiment with different models, prompts, and configurations instead of being locked into one provider.

Reduced dependency

The project can run its AI functionality locally after the required model and dependencies have been downloaded.

Accessibility

Local inference avoids mandatory per-request charges from a proprietary AI API, although users still need suitable hardware and storage.

For a project built around preserving memories, I believe the technology should support ownership rather than create another dependency.

The Spirit of It

I kept coming back to one thought while building ECHOEVER:

Sometimes, what we miss most isn't a person or a place. It's a moment.

A photograph, a conversation, a family tradition, a random afternoon that we didn't realize would become a memory.

Technology cannot bring those moments back. But perhaps it can help us preserve the stories attached to them.

With ECHOEVER, I wanted to explore how open-source AI could help us keep those stories accessible, personal, and alive.

Some memories fade. Some echoes last forever.

Prize Categories

best use of gemma

Thank You

ReCode — Never Let Your LeetCode Progress Go to Waste ♻️

Nidhi Jha — Sun, 25 Oct 2026 16:00:15 +0000

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

What I Built

Solving a coding problem once doesn't mean you'll remember how to solve it a month later.

As a student preparing for software engineering internships, I've solved many LeetCode problems, but I realized that solving problems is only half the journey — revisiting them at the right time is what helps you retain the patterns.

That's why I came up with ReCode, a personal DSA revision system that helps developers turn their solved problems into a structured, repeatable revision routine.

Instead of manually tracking problems in spreadsheets or wondering what to revise next, ReCode is designed to help you:

📅 Schedule revisions: Revisit problems at configurable intervals, with a default schedule of Day 1, 3, 7, and 14.

🧠 Revise smarter: Record whether a problem felt easy, was a struggle, or had been forgotten, helping identify areas that need more practice.

📊 Track progress: Organize problems by topic and monitor your revision progress.

📥 Import existing progress: Bring in your previously solved problems through CSV or Excel files.

🔗 Return to the problem: Open the original LeetCode problem directly from your revision list.

🤖 Get AI-assisted revision support: Use an open-weight AI model to generate recall prompts and assist with revision.

The goal is simple: spend less time organizing your preparation and more time actually improving your problem-solving skills.

Demo

Code

Nidhijha-1001 / ReCode

ReCode

A voxel-inspired DSA study world for saving problems, revisiting them on a real schedule, and learning from your own recall.

ReCode turns spaced repetition into a small adventure: your saved problems become inventory, reviews become quests, and progress is earned from completed study—not simulated gameplay. Its forest-world artwork and block-built interface are original CSS illustrations; ReCode is not affiliated with Minecraft or Mojang.

Screenshots show the app running locally with its development sample problems. They are genuine captures of the working UI, not a static mockup.

At a glance

Study dashboard: real problem, due/overdue, topic, difficulty, weekly activity, and player-profile data.
Problem Bank: search and sort saved problems, filter by topic, difficulty, and revision state, inspect a selected problem, and open its LeetCode page.
Revision quests: rate recall as Easy, Struggled, or Forgot; completion updates the schedule, review history, XP, streak, and milestones.
Solve workflow: record…

View on GitHub

How I Built It

I wanted ReCode to be more than a basic AI wrapper. Its core value comes from a reliable revision workflow, with AI serving as an additional layer of assistance.

The planned technology stack includes:

Frontend: React and TypeScript for a responsive, interactive web experience.

Backend: FastAPI to handle application logic and API requests.

Database: SQLite to persist problem details, revision schedules, and progress.

Open-source AI: Gemma, run locally using Ollama, for AI-assisted recall and revision support.

One important design decision is that the revision schedule is deterministic. The application can calculate when a problem is due without asking an AI model to decide every revision date. This keeps the core workflow predictable, while AI is used where it adds value: helping users recall concepts and revisit their reasoning.

The aim is to build something that is useful as a personal tool today and has a clear path toward becoming a production-ready application.

Why Does Open Innovation Matter?

For a project like ReCode, open innovation isn't just about using a different model. It's about having more control over how the technology works.

Using an open-weight model such as Gemma through local inference makes experimentation with AI-assisted revision possible without making a proprietary AI API the foundation of the entire application.

It also opens up opportunities to:

Experiment with prompts and model behavior.

Run inference locally rather than depending entirely on an external AI service.

Keep the core scheduling and tracking functionality independent of any particular model.

Adapt the project as open-source models and tools improve.

I want ReCode to demonstrate that useful AI applications don't always need to depend on closed APIs. Developers can combine open models with conventional software engineering to build tools that solve real problems.

My Agent Session

The Lazy Payoff

Imagine having hundreds of solved problems but no reliable way to decide which ones deserve another look.

Without a revision system, you have to remember problem names, search through old submissions, maintain a spreadsheet, and decide what to practise each day.

ReCode aims to reduce that overhead by bringing revision scheduling, progress tracking, and AI-assisted recall into one place.

The payoff isn't solving fewer problems. It's getting more long-term value out of the problems you've already solved.

I built this idea around a problem I've experienced in my own DSA preparation, and I want to make consistent revision easier for other students facing the same challenge.

Prize Categories

best use of gemma and githubcopilot.

ReCode is my attempt to make DSA preparation a little more systematic: solve a problem, revisit it at the right time, and build knowledge that actually sticks.

Because progress isn't just about how many problems you solve. It's about how many you can still solve when it matters.

Hacktoberfest #OpenSource #AI #Gemma #DSA #WebDevelopment

BillBuddhi-Local-first AI invoice assistant for Indian small businesses using Gemma, Ollama, Tesseract OCR, and Streamlit.

Nidhi Jha — Mon, 19 Oct 2026 05:57:34 +0000

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

What I Built

BillBuddhi is a local-first AI invoice assistant designed for Indian freelancers and small businesses.
Manually entering invoice details into spreadsheets is tedious. BillBuddhi helps simplify this workflow by extracting invoice information, allowing users to review the results, flagging basic inconsistencies, and exporting reviewed invoices to a CSV ledger.
The goal is to help small businesses spend less time on manual data entry while keeping control over their financial records.

Demo

Watch the BillBuddhi demo:

The demo shows invoice upload, local AI extraction with Gemma, validation warnings, and CSV ledger export.

Code

Nidhijha-1001 / billbuddhi

BillBuddhi

A local-first AI invoice assistant for Indian freelancers and small businesses.

BillBuddhi turns invoice images and text-based PDFs into a reviewable ledger. It combines local OCR and PDF text extraction with structured extraction from Gemma running through Ollama. You can correct the extracted fields, check basic consistency flags, and export reviewed entries to CSV.

Prototype, not accounting software. AI output can be wrong. Review every value against the source document before recording it. BillBuddhi does not establish GST compliance or provide tax advice.

Features

Upload PNG, JPG/JPEG, or WEBP invoice images, or PDFs that contain selectable text.
Extract image text with Tesseract OCR and PDF text with pypdf.
Use a local Gemma model via Ollama to suggest structured invoice fields
- Supplier name, invoice number, invoice date, and currency
- Subtotal, CGST, SGST, IGST, other charges, and total
- Optional notes
Review and manually correct the suggested values before adding an invoice.
…

View on GitHub

How I Built It

I built BillBuddhi using Python and Streamlit, with Tesseract OCR for invoice images and pypdf for text-based PDFs.
Gemma runs locally through Ollama and extracts structured invoice fields. Python independently checks basic arithmetic discrepancies, while users can review and correct extracted values before adding invoices to the ledger.
The application supports CSV export for further bookkeeping.

Why Does Open Innovation Matter?

Financial documents can contain sensitive business information. Running an open-weight model locally gives users greater control over where their invoice data is processed.
Using Gemma through Ollama also reduces dependence on hosted AI APIs and their per-request pricing. Developers can experiment with different compatible models and improve the extraction workflow without committing to a single proprietary AI provider.
Local inference still requires suitable hardware and electricity, and extraction errors remain possible. BillBuddhi therefore includes human review rather than treating AI output as automatically correct.

Prize Categories

Best Use of Gemma

CrackLife: Discover the Hidden Ecosystems of Your City 🌱

Nidhi Jha — Sun, 11 Oct 2026 17:54:34 +0000

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

What I Built

CrackLife is a prototype of an AI-powered urban nature field guide designed to help people step away from their screens and explore the small ecosystems around them.

Nature isn't limited to forests and national parks. Moss growing between pavement cracks, plants emerging through concrete, and pollinators visiting urban gardens all deserve attention.

CrackLife aims to turn a short walk into a nature-observation quest, encouraging users to notice their surroundings and record what they discover.

Demo

Code

Nidhijha-1001 / Cracklife_project

CrackLife

CrackLife is a local-first AI field guide for noticing the tiny ecosystems hiding in cities. It turns a short walk into a structured, safe observation quest and keeps a private field journal on the user's machine.

Why this project is differentiated

Instead of a generic chatbot, CrackLife turns the AI into a guided, real-world discovery tool:

prompts users for duration, location type, difficulty, and biodiversity focus
generates three concrete observation tasks plus a reflection prompt
encourages careful, respectful urban observation without disturbing wildlife
saves a private journal locally so the experience remains useful after the walk ends

Architecture

Frontend: Streamlit UI
Core logic: Python app with quest generation and journal persistence
AI layer: Ollama with an open-weight model such as Gemma or Llama, if installed locally
Fallback behavior: transparent heuristic prompts when an open model is unavailable
Storage: local JSON journal in the project directory

Open-source AI rationale

The…

View on GitHub

How I Built It

I built the prototype with Python and Streamlit. The project includes an interactive interface for generating nature quests and a field-journal workflow.

The project is designed to integrate an open-weight language model through local inference. However, that integration still needs to be tested before I can claim that the current version generates AI-powered quests.

Why Does Open Innovation Matter?

Open-source software and open-weight AI models can give developers greater control over how applications are built and adapted.

For a project like CrackLife, local inference could eventually reduce reliance on external AI services and help keep users' observations on their own devices. Open model access could also make it easier to experiment with different models and adapt the experience for different environments.

These are goals for the project rather than verified capabilities of the current prototype.

new to the community

Nidhi Jha — Sat, 10 Oct 2026 14:32:13 +0000

Hii, I m new to this community. Happy to be here.

DEV Community: Nidhi Jha

ECHOEVER: Every Memory Deserves an Eternity | Building a Digital Memory Archive with Open-Source AI

What I Built

Demo

Code

Nidhijha-1001 / memento

ECHOEVER

Every Memory Deserves an Eternity.

Contents

The idea

How I Built It

Why Does Open Innovation Matter?

The Spirit of It

Prize Categories

ReCode — Never Let Your LeetCode Progress Go to Waste ♻️

What I Built

Demo

Code

Nidhijha-1001 / ReCode

ReCode

At a glance

How I Built It

Why Does Open Innovation Matter?

My Agent Session

The Lazy Payoff

Prize Categories

Hacktoberfest #OpenSource #AI #Gemma #DSA #WebDevelopment

BillBuddhi-Local-first AI invoice assistant for Indian small businesses using Gemma, Ollama, Tesseract OCR, and Streamlit.

What I Built

Demo

Code

Nidhijha-1001 / billbuddhi

BillBuddhi

Features

How I Built It

Why Does Open Innovation Matter?

Prize Categories

CrackLife: Discover the Hidden Ecosystems of Your City 🌱

What I Built

Demo

Code

Nidhijha-1001 / Cracklife_project

CrackLife

Why this project is differentiated

Architecture

Open-source AI rationale

How I Built It

Why Does Open Innovation Matter?

new to the community