DEV Community: jaya patel

Her Voice - The Ghost in the Machine That Listens Without Judging

jaya patel — Sun, 01 Nov 2026 05:59:53 +0000

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

Team Submission: Developed in equal partnership by @https://github.com/jaya030888 & @https://github.com/mdnawedullah09.

👥 Teamwork & Collaborative Contribution

This project was built from the ground up as a balanced, equal team collaboration between two developers. Every component of Her Voice was designed, implemented, and verified together across frontend and backend boundaries:

Developer 1 (@jaya patel):
- AI Architecture & Fallback Engine: Engineered the local Ollama LLM integration (Llama 3 / Mistral) and the deterministic trauma-informed RAG retrieval engine with legal citations (PWDVA 2005, POSH Act, BNS).
- Backend API & Data Models: Designed Express routes, Prisma schema, server-side PII moderation middleware, and automated unit testing suites with Vitest and Supertest.
- Crisis Triage Engine: Architected the multi-stage Quick Report crisis triage flow and emergency escalation pathways.
Developer 2 (@Nawed Ullah):
- Frontend Experience & Design System: Crafted the accessible, responsive React 18 & Tailwind UI, high-contrast typography, and calming trauma-informed aesthetic.
- Memorial & NGO Action: Built the unified Remember Her, Change Tomorrow memorial with virtual candle tributes and direct donation pledges to verified grassroots NGOs.
- Community & Safety Controls: Developed HerSpace pseudonym generation, empathetic reactions, interactive comment flows, and the instant Quick Exit escape system.

What We Built

What if the ghost inside the machine wasn't something to fear, but something that listened when nobody else would?

Her Voice is an open-source, trauma-informed digital platform built to protect, support, educate, and empower women facing domestic abuse, workplace harassment, digital stalking, and gender-based violence.

Inspired by the Halloween and Día de los Muertos themes, we wanted to confront a different kind of horror. The real horror isn't always supernatural. Often, it exists in silence, coercive control, isolation, and the voices society ignores or silences.

Her Voice reimagines this concept into a sanctuary where technology acts as an empathetic shield rather than a surveillance mechanism.

Key Pillars of Her Voice

💜 1. HerSpace — Anonymous Solidarity Community

A pseudonymous safe space where women share personal experiences and find community support without exposing their real identities.
Features automatic pseudonym allocation (e.g. SilverLily88, CourageWing42).
Filterable across six recovery categories: Domestic Abuse, Harassment, Sexual Violence, Relationships, Discrimination, and Recovery.
Empathetic reactions (Solidarity, Hug, Heart), supportive commenting threads, and private bookmarks.
Client-side and server-side PII regex scanning that flags accidental phone numbers, emails, or physical addresses before publication.

🛡️ 2. Quick Report — Guided Crisis Triage

A five-step structured crisis workflow designed to help users document incidents calmly and methodically:
1. Immediate danger self-assessment.
2. Incident category and narrative documentation.
3. Timeline & evidence logging (with simulated voice-note preservation).
4. Explicit-consent city-level location sharing for localized help.
5. Actionable pathways to emergency helplines, legal aid, and cybercrime reporting.
Direct one-tap connections to India's 112 universal emergency line and 1091 / 181 Women Helplines.
Note: Reports remain strictly internal private drafts and are never transmitted to authorities without survivor consent.

🌱 3. SafePath — Practical Safety Hub

Concrete, action-oriented survival tools for women in crisis:
- Emergency Grab-Bag Checklist: Vital documents, cash, medicines, emergency keys, and digital backups.
- Secret Code-Word Protocol: Discreet distress phrases to alert trusted friends without alerting abusers.
- 8-Question Relationship Red-Flag Reflection: Objective assessment of coercive control and emotional manipulation.
- Verified Helpline Directory: Direct access to Sakhi One Stop Centres, District Legal Services Authorities (DLSA), and NCW helplines.

🤖 4. Saathi AI — Support Companion (साथी AI)

The technical heart of our Ghost in the Machine concept: a privacy-first conversational companion powered by open-weight models through Ollama (Llama 3, Mistral, Qwen 2.5).
Zero Commercial API Dependency: Operates entirely with local open weights or on-device deterministic RAG.
Trauma-Informed Boundaries: Never claims to be a lawyer or therapist, never victim-blames, and never validates abusive behavior.
Bilingual Support: Fluent in English, Hindi, and conversational Hinglish.
Graceful Fallback: If Ollama or local LLM instances are offline, Saathi transitions to a deterministic RAG knowledge base covering Indian legal frameworks (PWDVA 2005, POSH Act, BNS provisions) and verified helplines.

🕯️ 5. Remember Her, Change Tomorrow & NGO Connect

A digital memorial honoring historic women whose lives sparked movements for reform (Jyoti Singh / Nirbhaya, Bhanwari Devi, Mahsa Amini).
Visitors can light digital candles of remembrance and read their stories.
Transform Tribute into Action: Seamlessly connects memorial stories to verified Indian grassroots NGOs (Sneha India, Majlis Legal Centre, Jagori, Sayfty, Breakthrough India) with direct links and solidarity donation pledges.

🔐 6. Privacy & Digital Safety Controls

Quick Exit (ESC): Instantly redirects the browser to Google Search and clears temporary inputs.
Discreet Camouflage Mode: Toggles the screen into a generic weather dashboard.
Zero Tracking: No advertising cookies or third-party behavioral trackers.

Demo

Live Platform: https://lastweek-three.vercel.app/
Interactive Features: Test out Saathi AI, explore the verified safety checklists on SafePath, read community stories in HerSpace, and light a tribute candle.

Code

jaya030888 / her-voice

Her Voice 💜

“Your story matters. Your identity is yours to protect.”

Her Voice is a trauma-informed, production-ready full-stack web application dedicated to women’s safety, support, legal literacy, and community awareness. The platform brings together an anonymous peer-to-peer solidarity community, streamlined crisis triage, practical safety planning tools, an open-weight local AI companion (Saathi AI), interactive educational branching scenarios (The Reality Simulator), a digital memorial honoring landmark legal reforms (Remember Her, Change Tomorrow), and a searchable directory of verified NGOs.

🎃 Hackathon Theme: Halloween & Día de los Muertos

Inspired by the Hacktober / Día de los Muertos hackathon themes, the central ethos of Her Voice is that real-world violence, coercive control, and victim-blaming are vastly more terrifying than fictional monsters or haunted houses.

Drawing from Día de los Muertos traditions, remembrance is not passive grief—it is active commemoration through systemic awakening, legal literacy,…

View on GitHub

The complete codebase is 100% open-source under the MIT License. Repository: https://github.com/jaya030888/her-voice

How We Built It

We wanted to build an architecture where artificial intelligence wasn't an opaque third-party black box, but a transparent, reproducible, privacy-first companion.

       ┌───────────────────────────────┐
       │     React 18 + Vite SPA       │
       │    (HerSpace / SafePath)      │
       └──────────────┬────────────────┘
                      │ (REST /api)
                      ▼
       ┌───────────────────────────────┐
       │     Express + TypeScript      │
       │  (Helmet, Rate-Limits, CORS)  │
       └───────┬──────────────┬────────┘
               │              │
       ┌───────▼───────┐  ┌───▼────────────────────────┐
       │  Prisma ORM   │  │    Saathi AI Engine        │
       │ (SQLite/PG)   │  │ ┌────────────────────────┐ │
       └───────────────┘  │ │ Ollama (Llama 3 Local) │ │
                          │ └───────────┬────────────┘ │
                          │             │ (fallback)   │
                          │ ┌───────────▼────────────┐ │
                          │ │ Verified RAG & Legal KB│ │
                          │ └────────────────────────┘ │
                          └────────────────────────────┘

The AI Architecture Behind Saathi AI

Local Inference: Ollama daemon hosting open-weight models (llama3:latest, mistral, qwen2.5).
Deterministic RAG Pipeline: Curated dataset of Indian legal rights, procedural safety protocols, and emergency helpline networks.
Safety Guardrails: Strict regex pre-filters for physical danger indicators (triggering emergency 112/181 alerts) and zero victim-blaming prompt constraints.
Graceful Degradation: Serverless-friendly design ensuring users still receive accurate legal and safety guidance even when local Ollama inference is offline.

Full Technology Stack

Frontend: React 18, Vite 5, TypeScript, Tailwind CSS, Framer Motion, Lucide React.
Backend: Node.js, Express.js, TypeScript, Prisma ORM, SQLite / PostgreSQL.
Security: Helmet, Express Rate Limiter, Zod schema validation, PII scrubbing.
Testing & CI/CD: Vitest, Supertest, GitHub Actions CI matrix.
Deployment: Configured for single-domain multi-service orchestration via Vercel Services (vercel.json).

Why Does Open Innovation Matter?

Privacy is Not a Premium Luxury: Survivors discussing sensitive trauma should never have their words harvested for commercial advertising or sent to closed proprietary clouds. Open weights keep data sovereign.
Accessible to Grassroots Non-Profits: Small NGOs and shelters cannot afford recurring per-token commercial API costs. Open-source models allow non-profits to host resilient guidance tools affordably.
Auditability & Algorithmic Trust: In sensitive domains like legal aid and domestic violence, opaque black boxes can cause harm. Open innovation allows advocates, survivors, and legal scholars to inspect every prompt and rule.
Community Co-Creation: Anyone can contribute legal updates, local helpline numbers, translated regional dialects, and trauma-informed improvements.

The Spirit of It 💜

Building Her Voice brought us back to one central thought:

What happens when someone finally finds a place where they can speak without fear of judgment?

For anyone enduring abuse or coercive control, seeking help can be the most terrifying step of all. We built Her Voice so that nobody has to surrender their identity before receiving guidance, nobody has their credibility questioned, and nobody is forced to speak before they are ready.

Saathi AI is the ghost in the machine: an invisible ally that stays awake through the darkest hours, listens with compassion, and reminds survivors that they are never truly alone.

And through the memorial, we remember that honoring the women who were silenced must inspire real-world action to protect those who are still here.

Team Credits

Developer 1: @jaya patel — AI Architecture, Backend APIs, Crisis Triage & Tests
Developer 2: @Nawed Ullah — Frontend Architecture, UI/UX, Memorial & Community Features

A woman's voice should never have to disappear for the world to finally listen. 💜

devchallenge #hf26challenge #hacktoberfest #opensource #ai

Hydroniq

jaya patel — Mon, 26 Oct 2026 05:59:15 +0000

💧 Hydroniq: Making Water Problem Reporting Simpler with Open-Source AI

We built Hydroniq, a water intelligence platform designed to make community water problem reporting and tracking easier. Our goal is to combine open-source AI with a practical social and environmental challenge: reducing water wastage and making water-related problems easier to organize and monitor.

What I Built

Hydroniq helps organize reports about leaking public taps, broken handpumps, and other water infrastructure problems in one place.

Reporting these problems is only the first step. Organizing complaints, summarizing their details, identifying missing information, and tracking their progress can also involve repetitive administrative work.

Hydroniq aims to simplify that workflow.

Key Features

Community Reporting: Submit water-related problems with descriptions and location details.
Report Tracking: Track reports through Pending, In Progress, and Resolved statuses.
Live Dashboard: View total submissions, pending reports, and resolved issues.
Impact Analytics: Explore which types of water problems are reported most frequently.
Area Monitoring: Organize reports by state, district, and location.
AI-Assisted Analysis: Use Gemma to suggest issue categories, summarize reports, identify missing information, and recommend next steps for human review.

The idea behind Hydroniq is simple: reduce repetitive work so people can spend more time understanding problems and deciding what needs attention.

AI-generated analysis is only a suggestion. It does not independently verify a complaint or replace human decision-making.

Demo

🌐 Live Demo: https://hydroniq-theta.vercel.app

The deployed application connects a React frontend to a FastAPI backend and a cloud-hosted MySQL database. Reports can be submitted, stored, retrieved, and displayed on the dashboard.

Current limitation: The AI-assisted analysis works with a local Ollama and Gemma setup during development. Hosted AI inference is still a future improvement.

Code

💻 GitHub Repository: https://github.com/mdnawedullah09/Hydroniq

Hydroniq is open source, and we welcome feedback, ideas, and contributions from the community.

How I Built It

We used the following technology stack:

Frontend: React and Vite
Backend: Python and FastAPI
Database: MySQL hosted on Aiven
Open-Source AI: Gemma 3 4B through Ollama
Frontend Deployment: Vercel
Backend Deployment: Render
Version Control: Git and GitHub

How the AI Workflow Works

A user provides a water problem report, including its issue type, location, and description.
The report information is sent to the Gemma model through Ollama in the local development setup.
The model suggests a category, generates a short summary, highlights missing information, and proposes a practical next step.
A human reviewer can use these suggestions to help decide what to examine next.

The model is not treated as an authority. Human review remains important before any official action is taken.

Building Hydroniq also helped us learn how a frontend, backend APIs, a persistent cloud database, and a local AI model can fit together in a full-stack application.

Why Does Open Innovation Matter?

Open innovation matters because developers should have the freedom to experiment with AI, understand its limitations, and adapt it to real-world needs.

Using the open-weight Gemma model through Ollama allowed us to explore AI-assisted report analysis locally instead of making a closed AI API the core of our workflow.

This gave us more control over experimentation and prompt design, while helping us understand the practical considerations of running a model locally.

For a project focused on community welfare and environmental sustainability, accessible and adaptable AI tools can help developers build solutions around local problems.

Open-source technology also makes it easier for others to inspect the code, learn from the implementation, and contribute improvements.

My Agent Session

We do not have a DevRelay agent session to share yet.

The Lazy Payoff

Hydroniq is designed to reduce the repetitive steps involved in organizing water problem reports.

Instead of manually preparing an initial summary for each complaint, AI-assisted analysis can suggest a category, summarize the description, flag missing details, and recommend a next step for a human reviewer.

The reporting, tracking, dashboard, analytics, and database workflow has been implemented. However, we have not yet measured time savings with real community users, so we will not claim a specific number of minutes saved.

Our next goal is to test the workflow with real users and measure whether it reduces the effort required to review and organize reports.

For us, “Slow and Steady” means building a useful foundation first, testing it honestly, and improving it step by step rather than making unsupported claims about impact.

Prize Categories

Our project focuses on:

Open-source AI
Practical AI-assisted workflow automation
Social impact and environmental sustainability

We will choose the official partner prize categories that match the challenge rules before submitting.

Team

We worked together on Hydroniq, combining our ideas and development efforts to build a water intelligence platform that uses open-source AI to support community water problem reporting and water conservation.

MD Nawedullah Saif — Co-Developer
Jaya Patel — Co-Developer

We collaborated on the project and shared responsibility for bringing Hydroniq from an idea to a working application, combining full-stack development with our goal of supporting communities and protecting water resources.

Final Thoughts

Hydroniq started with a simple idea: make water-related problems easier to report, organize, and monitor.

This project has been a learning experience in full-stack development, deployment, database integration, and experimenting with open-source AI in a practical workflow.

We hope to keep improving Hydroniq, make its AI-assisted features more accessible, and explore how technology can help communities protect an essential resource: water.

We would love to hear your feedback and suggestions. What features would make a community water reporting platform more useful?

Hacktoberfest #OpenSource #AI #WaterConservation

My Garden World

jaya patel — Sun, 11 Oct 2026 17:12:28 +0000

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

What I Built

My Garden World — Discover nature. Grow with intelligence. Watch life unfold.

My Garden World is an AI-powered nature companion designed to encourage people to step away from their screens and reconnect with the natural living world.

It brings together three core experiences:

AI-Powered Plant Discovery: Identify real-world flora on-device and explore botanical facts, sunlight, watering needs, and ecological importance.
AI-Powered Garden Care & Vital Diagnostics: Get species-tailored treatment plans, NASA air-filtering scores, ASPCA pet-safety advisories, and photo growth tracking.
Virtual Garden Builder & Nature Quests: Build and customize an interactive 8×8 sanctuary grid, complete outdoor mindfulness quests, and earn Flora Coins and XP.

The goal is to make learning about plants and gardening interactive, accessible, and enjoyable for beginners and nature enthusiasts alike.

Demo

Live Demo: https://my-garden-world-tau.vercel.app/

Explore the application and discover how technology can make outdoor exploration more engaging.

Code

GitHub Repository: https://github.com/jaya030888/my-garden-world

How I Built It

I built My Garden World as a modern, local-first full-stack application:

Frontend: React 19, TypeScript, Vite, and Tailwind CSS deployed on Vercel.
On-Device Vision AI: Client-side Vision Transformer (google/vit-base-patch16-224) executed locally via Transformers.js and ONNX Runtime Web (WebAssembly).
Backend: Node.js and Express REST API with Helmet security and rate limiting.
Database: Serverless PostgreSQL hosted on Neon, with Prisma ORM for schema migrations and relational data management.
Local Persistence: Dexie.js (IndexedDB) caching for instant offline state resilience.

Why Does Open Innovation Matter?

For me, open innovation means having the freedom to build AI that serves people and nature—not being constrained by proprietary, black-box APIs.

With My Garden World, open-weight AI unlocks fundamental advantages:

Complete User Privacy Outdoors: Because the Vision Transformer executes directly in the user's browser sandbox via WebAssembly, no plant photographs ever leave the user's device.
Zero Proprietary Cloud Tax: We avoid expensive per-token closed-source vision APIs. The intelligence belongs to the user, running on their own phone or laptop at $0 inference cost.
Independence and Adaptability: Instead of being locked into a single provider's terms or pricing, the community can inspect, benchmark, and swap open-source vision architectures (ViT, MobileNet, Swin) freely.

The philosophy is simple: use open technology not to keep us glued to screens, but to inspire us to step outside, observe plants, and build environmental awareness.

Taking It Outside

The true test of a nature application is how it performs beyond the desk.

My next step is taking My Garden World out on trails and city parks to test identification across varying daylight conditions, diverse wild flora, and different growth stages.

The goal isn't just an interface that looks nice on a screen, but an experience that sparks genuine curiosity about the living world outside.

Challenges I Faced

Running Neural Models in the Browser: Compiling and running an 86.6M parameter Vision Transformer on-device required balancing model size and latency. Using INT8 quantization through Transformers.js allowed the model to initialize and classify within milliseconds.
Decoupling Container Artifacts: Real-world plant photos frequently feature pots, planters, and vases. We built a botanical taxonomy filter to recognize and decouple container objects rather than hallucinating false species classifications.
State Synchronization: Harmonizing instant client-side Dexie IndexedDB caching with cloud PostgreSQL persistence on Neon to keep user gardens, coins, and plant journals synchronized.

What's Next?

Expanding the botanical taxonomy database with more regional native flora.
Introducing seasonal planting calendars and microclimate weather tracking.
Adding community-shared local nature walks and biodiversity quests.
Further optimizing WebGPU inference for even faster mobile scans.

My Agent Session

I paired with an AI coding assistant to architect the full-stack system, evaluate open-weight vision transformers, and debug TypeScript and deployment configurations.

Prize Categories

Open-Source AI
Touch Grass