DEV Community: Yug Patel

NecroShell & OfrendaAI: The Open-Weight Haunted Terminal & Digital Remembrance Shrine

Yug Patel — Sun, 01 Nov 2026 04:16:07 +0000

Proprietary APIs get deprecated, price-hiked, or blocked by firewalls. An open-weight model belongs to humanity forever. If the power grid falters on Halloween night, NecroShell still boots from local silicon with the lights out and the Wi-Fi down.

Freedom from Sanitized Cultural Bias: Día de los Muertos is a profound celebration where death is embraced with warmth, humor, sugar skulls, and Calaveritas Literarias. Closed cloud APIs frequently over-moderate terms like "death", "cemetery", "bones", or "ghosts". Open-weight models allowed us to prompt freely for rich, authentic, culturally resonant folk poetry without arbitrary safety filter refusals. --- ## The Spirit of It During testing, I inscribed a tribute to my late grandmother, mentioning how she always had fresh pan dulce and cinnamon coffee waiting whenever we arrived. When the open-weight model returned the rhyming Calaverita Literaria with golden marigold petals fluttering across the glowing CRT screen: > "La Catrina rondaba buscando una estrella, > y entre mil resplandores halló la más bella..." It gave me literal goosebumps. Technology often feels cold, disposable, and obsessed with the next sprint. Using open-source AI to build something that honors those who came before us—whether pioneers of computation or our own ancestors—reminded me of why we build software in the first place: to connect human souls across time. --- ## Prize Categories I am entering this submission into the following categories:
Best Use of Gemma: Powered by Google's Gemma 2 open-weight model (google/gemma-2-2b-it) for the séance persona engine and bilingual Día de los Muertos calaverita generation.
Best Use of Render: Includes a production render.yaml blueprint and Dockerfile for instantaneous 1-click cloud deployment.
Best Use of ElevenLabs: Integrates ElevenLabs' text-to-speech API to voice the eerie whispers and séances aloud.
Best Use of MongoDB Atlas: Employs an exportable document schema for persistent mementos and tributes on the digital altar.

SlothDesk – Let AI Handle the Boring Stuff

Yug Patel — Fri, 30 Oct 2026 13:29:33 +0000

🦥 SlothDesk – Let AI Handle the Boring Stuff

This is a submission for the "Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady" (https://dev.to/challenges/hacktoberfest-week3-2026-10-19)

What I Built

Meet SlothDesk, a local-first AI productivity assistant designed for people who want to spend less time managing files and more time doing meaningful work.

We all have those repetitive tasks we keep postponing: organizing downloads, renaming documents, understanding lengthy PDFs, and sorting files.

SlothDesk aims to handle these boring tasks with the help of local AI.

Its goal is simple: Work smarter, automate the boring stuff, and embrace the slow-and-steady lifestyle.

Key Features

AI-powered document summarization.
Automatic file organization.
Intelligent file renaming.
Local PDF analysis.
A simple dashboard to manage everyday tasks.

Demo

[Add your demo video or application link here.]

Code

GitHub Repository: [Add your repository link here.]

How I Built It

SlothDesk is built around open-source AI and local inference.

My technology stack:

Python: Core application logic and automation.
Ollama: Running AI models locally.
Qwen: Open-weight language model for intelligent text processing.
Streamlit: Interactive user interface.
SQLite: Lightweight local database.
PyMuPDF: PDF text extraction and processing.

Instead of sending documents to an external AI API, SlothDesk uses locally hosted models through Ollama.

The application connects AI capabilities with practical Python automation to reduce repetitive manual work.

Why Does Open Innovation Matter?

For me, open innovation is about having control over the technology we build.

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

Privacy: Personal documents can be processed locally without uploading their contents to an external AI provider.
Independence: The application does not depend on a paid proprietary AI API.
Flexibility: Models can be changed depending on available hardware and performance requirements.
Accessibility: Developers and students can experiment with AI without needing expensive API subscriptions.
Customization: We can modify the application's behaviour and automation logic according to our needs.

Open-source AI makes it possible to build useful tools that are accessible, customizable, and under our control.

The Lazy Payoff

The idea behind SlothDesk is to reduce the small repetitive tasks that consume our day.

Instead of manually opening documents, renaming files, and organizing folders, users can delegate supported tasks to their local AI assistant.

The ultimate goal is to spend less time on routine work and more time on learning, building, and creating.

Prize Categories

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

Move slowly. Automate wisely. Let AI do the boring work. 🦥

NecroShell & OfrendaAI: The Open-Weight Haunted Terminal & Digital Remembrance Shrine

Yug Patel — Thu, 29 Oct 2026 15:18:23 +0000

Freedom from Sanitized Cultural Bias: Día de los Muertos is a profound celebration where death is embraced with warmth, humor, sugar skulls, and Calaveritas Literarias. Closed cloud APIs frequently over-moderate terms like "death", "cemetery", "bones", or "ghosts". Open-weight models allowed us to prompt freely for rich, authentic, culturally resonant folk poetry without arbitrary safety filter refusals. --- ## The Spirit of It During testing, I inscribed a tribute to my late grandmother, mentioning how she always had fresh pan dulce and cinnamon coffee waiting whenever we arrived. When the open-weight model returned the rhyming Calaverita Literaria with golden marigold petals fluttering across the glowing CRT screen: > "La Catrina rondaba buscando una estrella, > y entre mil resplandores halló la más bella..." It gave me literal goosebumps. Technology often feels cold, disposable, and obsessed with the next sprint. Using open-source AI to build something that honors those who came before us—whether pioneers of computation or our own ancestors—reminded me of why we build software in the first place: to connect human souls across time. --- ## Prize Categories I am entering this submission into the following categories:
Best Use of Gemma: Powered by Google's Gemma 2 open-weight model (google/gemma-2-2b-it) for the séance persona engine and bilingual Día de los Muertos calaverita generation.
Best Use of Render: Includes a production render.yaml blueprint and Dockerfile for instantaneous 1-click cloud deployment.
Best Use of ElevenLabs: Integrates ElevenLabs' text-to-speech API to voice the eerie whispers and séances aloud.
Best Use of MongoDB Atlas: Employs an exportable document schema for persistent mementos and tributes on the digital altar.

Freedom from Sanitized Cultural

Yug Patel — Wed, 28 Oct 2026 16:57:34 +0000

Freedom from Sanitized Cultural Bias: Día de los Muertos is a profound celebration where death is embraced with warmth, humor, sugar skulls, and Calaveritas Literarias. Closed cloud APIs frequently over-moderate terms like "death", "cemetery", "bones", or "ghosts". Open-weight models allowed us to prompt freely for rich, authentic, culturally resonant folk poetry without arbitrary safety filter refusals. --- ## The Spirit of It During testing, I inscribed a tribute to my late grandmother, mentioning how she always had fresh pan dulce and cinnamon coffee waiting whenever we arrived. When the open-weight model returned the rhyming Calaverita Literaria with golden marigold petals fluttering across the glowing CRT screen: > "La Catrina rondaba buscando una estrella, > y entre mil resplandores halló la más bella..." It gave me literal goosebumps. Technology often feels cold, disposable, and obsessed with the next sprint. Using open-source AI to build something that honors those who came before us—whether pioneers of computation or our own ancestors—reminded me of why we build software in the first place: to connect human souls across time. --- ## Prize Categories I am entering this submission into the following categories:
Best Use of Gemma: Powered by Google's Gemma 2 open-weight model (google/gemma-2-2b-it) for the séance persona engine and bilingual Día de los Muertos calaverita generation.
Best Use of Render: Includes a production render.yaml blueprint and Dockerfile for instantaneous 1-click cloud deployment.
Best Use of ElevenLabs: Integrates ElevenLabs' text-to-speech API to voice the eerie whispers and séances aloud.
Best Use of MongoDB Atlas: Employs an exportable document schema for persistent mementos and tributes on the digital altar.

NecroShell & OfrendaAI: The Open-Weight Haunted Terminal & Digital Remembrance Shrine

Yug Patel — Wed, 28 Oct 2026 16:32:48 +0000

Freedom from Sanitized Cultural Bias: Día de los Muertos is a profound celebration where death is embraced with warmth, humor, sugar skulls, and Calaveritas Literarias. Closed cloud APIs frequently over-moderate terms like "death", "cemetery", "bones", or "ghosts". Open-weight models allowed us to prompt freely for rich, authentic, culturally resonant folk poetry without arbitrary safety filter refusals. --- ## The Spirit of It During testing, I inscribed a tribute to my late grandmother, mentioning how she always had fresh pan dulce and cinnamon coffee waiting whenever we arrived. When the open-weight model returned the rhyming Calaverita Literaria with golden marigold petals fluttering across the glowing CRT screen: > "La Catrina rondaba buscando una estrella, > y entre mil resplandores halló la más bella..." It gave me literal goosebumps. Technology often feels cold, disposable, and obsessed with the next sprint. Using open-source AI to build something that honors those who came before us—whether pioneers of computation or our own ancestors—reminded me of why we build software in the first place: to connect human souls across time. --- ## Prize Categories I am entering this submission into the following categories:
Best Use of Gemma: Powered by Google's Gemma 2 open-weight model (google/gemma-2-2b-it) for the séance persona engine and bilingual Día de los Muertos calaverita generation.
Best Use of Render: Includes a production render.yaml blueprint and Dockerfile for instantaneous 1-click cloud deployment.
Best Use of ElevenLabs: Integrates ElevenLabs' text-to-speech API to voice the eerie whispers and séances aloud.
Best Use of MongoDB Atlas: Employs an exportable document schema for persistent mementos and tributes on the digital altar.

NecroShell & OfrendaAI: The Open-Weight Haunted Terminal & Digital Remembrance Shrine

Yug Patel — Tue, 27 Oct 2026 13:45:28 +0000

Freedom from Sanitized Cultural Bias: Día de los Muertos is a profound celebration where death is embraced with warmth, humor, sugar skulls, and Calaveritas Literarias. Closed cloud APIs frequently over-moderate terms like "death", "cemetery", "bones", or "ghosts". Open-weight models allowed us to prompt freely for rich, authentic, culturally resonant folk poetry without arbitrary safety filter refusals. --- --- ## The Spirit of It During testing, I inscribed a tribute to my late grandmother, mentioning how she always had fresh pan dulce and cinnamon coffee waiting whenever we arrived. When the open-weight model returned the rhyming Calaverita Literaria with golden marigold petals fluttering across the glowing CRT screen: > "La Catrina rondaba buscando una estrella, > y entre mil resplandores halló la más bella..." It gave me literal goosebumps. Technology often feels cold, disposable, and obsessed with the next sprint. Using open-source AI to build something that honors those who came before us—whether pioneers of computation or our own ancestors—reminded me of why we build software in the first place: to connect human souls across time. --- ## Prize Categories I am entering this submission into the following categories:
Best Use of Gemma: Powered by Google's Gemma 2 open-weight model (google/gemma-2-2b-it) for the séance persona engine and bilingual Día de los Muertos calaverita generation.
Best Use of Render: Includes a production render.yaml blueprint and Dockerfile for instantaneous 1-click cloud deployment.
Best Use of ElevenLabs: Integrates ElevenLabs' text-to-speech API to voice the eerie whispers and séances aloud.
Best Use of MongoDB Atlas: Employs an exportable document schema for persistent mementos and tributes on the digital altar.

Unscatter AI: The Open-Source ADHD & Executive Dysfunction Copilot Built

Yug Patel — Tue, 13 Oct 2026 06:19:59 +0000

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

💖 The Inspiration: Who I Built This For

My close friend Kavya is an exceptionally talented developer and university student who lives with adult ADHD and severe executive dysfunction.

Watching someone you care about get trapped in "ADHD Task Paralysis" (the freeze state) is heartbreaking:

The Paralysis Spiral: When coursework, project deadlines, and messages accumulate, her brain enters cognitive overload. Traditional productivity apps (Notion, Todoist, Asana) make it worse: they demand heavy taxonomy, tag-sorting, and show a giant wall of 30 intimidating tasks that induces panic.
Rejection Sensitive Dysphoria (RSD): Short, unpunctuated messages like "We need to talk tomorrow morning" or "Check my comments ASAP" cause acute chest-tightening anxiety and dread, making her freeze before responding.
The Quiet Room Struggle: Working in isolation is mentally draining. ADHD psychology shows that "Body Doubling" (having a supportive, non-judgmental presence in the room) helps overcome initiation friction.

I decided to build Unscatter AI as an open-source gift for Kavya—and for anyone fighting executive paralysis.

🚀 What I Built

Unscatter AI is a local-first, distraction-free companion engineered to lower the activation energy of daily life:

1. 🌪️ Brain Dump Unscrambler ("Unfreeze Inertia")

Instead of sorting tasks into folders, users speak or type their raw stream-of-consciousness chaos into the box (e.g. "I'm sitting on the floor panicking, I have 8 assignments due, my inbox is full of scary emails, my desk is messy, I haven't eaten").

The AI decomposes the panic into bite-sized 2-to-5 minute atomic micro-steps.
It highlights a "🌱 Micro-Launch Step" requiring near-zero willpower (e.g. "Open editor tab and drink a sip of water") to get the dopamine flywheel turning.

2. 🥄 Spoon Theory Energy Calibrator

Calibrates tasks against current mental energy (from 1 Spoon = burnout survival, up to 5 Spoons = hyperfocus surge), eliminating guilt on low-energy days.

3. 🛡️ RSD Shield & Communication Calmer

Incoming Shield: Deconstructs blunt emails from bosses/professors into objective facts, performs a psychological reality check on why the sender isn't angry, and generates a calm 1-click reply.
Outgoing Softener: Transforms anxious, over-apologetic drafts ("I'm so so sorry for existing and being late...") into grounded, professional communication ("Thank you for your patience...").

4. ☕ Virtual Body Doubler with Web Audio Synthesizer

A quiet co-working companion providing parallel presence with timer check-ins.
Built-in Web Audio API ambient sound generator that synthesizes calming brown noise and rain frequencies directly in pure browser code without loading external MP3 files.
Visual breathing ring to down-regulate sympathetic nervous system arousal.

5. 🏆 Friend's Compassion Vault (Evidence Locker)

ADHD brains often experience "emotional amnesia"—when frozen, they forget every prior achievement. This vault stores concrete evidence of competence and encouraging notes from friends.

🧠 Why Open-Source AI was Integral to This Project

The Hacktoberfest challenge specifically calls out the necessity of an open-source AI approach. For Unscatter AI, open-weight models were essential for three non-negotiable reasons:

1. Radical Privacy for Vulnerable Mental Health Data

A brain dump often contains intensely private confessions: emotional distress, struggles with medication, workplace anxiety, and personal diaries. Routing these thoughts through commercial cloud LLM APIs means handing sensitive psychological telemetry to advertising giants. Running open-weight models (via Ollama llama3.2 or quantized edge models) ensures that every thought stays strictly on local silicon.

2. Zero-Cost Accessibility for Neurodivergent People

ADHD and executive dysfunction disproportionately cause financial precarity (the "ADHD tax"). Tools designed to help people function should never be locked behind \$20/month SaaS paywalls. Open-weight models ensure that anyone with a laptop can run their personal companion for free forever.

3. Empathy Tuning & Tone Freedom

Proprietary cloud models often respond with sterile, patronizing corporate cheerleading. With open weights, prompts and models are shaped to act like a gentle, realistic best friend who validates struggles rather than lecturing.

🛠️ Tech Stack & Architecture

   ┌─────────────────────────────────────────────────────────┐
   │             Responsive Web Interface                    │
   │  (Tailwind CSS, Phosphor Icons, Web Audio Synthesizer)  │
   └────────────────────────────┬────────────────────────────┘
                                │ HTTP / JSON API
   ┌────────────────────────────▼────────────────────────────┐
   │                     FastAPI Backend                     │
   │                  (Python 3.12, Uvicorn)                 │
   └─────────────┬─────────────────────────────┬─────────────┘
                 │                             │
   ┌─────────────▼─────────────┐ ┌─────────────▼─────────────┐
   │   Open-Source AI Engine   │ │     Local SQLite DB       │
   │ ├─ Ollama (Local Llama 3.2│ │     (unscatter.db)        │
   │ └─ Offline Edge Heuristics│ └───────────────────────────┘
   └───────────────────────────┘

Backend: Python 3.12, FastAPI, Uvicorn, Pydantic, HTTPX
AI Core: Ollama local inference (llama3.2, mistral, gemma2) with fallback to Edge Neurodivergent Heuristics
Frontend: Vanilla HTML5/JS, Tailwind CSS, Web Audio API, Web Speech Recognition API
Persistence: Local SQLite (unscatter.db)

💻 Source Code & Quickstart

The repository is open source and ready to run locally:

# 1. Clone the repository
git clone https://github.com/your-username/unscatter-ai.git
cd unscatter-ai

# 2. Install dependencies & run (or double-click run.bat on Windows)
cd backend
python -m pip install -r requirements.txt
python main.py

Then visit http://localhost:8001 in your browser.

🌟 Reflections & Next Steps

Building this project taught me that the best applications aren't built for abstract demographics—they are built for specific people you care about. Seeing my friend Kavya take a deep breath, break through a terrifying backlog in 10 minutes, and smile was worth every line of code.

Future plans include:

Expanding the offline Web Audio ambient engine with binaural beats
Native desktop tray notifications for body-double check-ins
Multi-device local P2P sync via WebRTC

BioTrail AI: The Privacy-First Offline Field Botanist

Yug Patel — Sun, 11 Oct 2026 04:25:38 +0000

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

published: true
tags: hacktoberfest, ai, opensource, python, webdev
cover_image: https://images.unsplash.com/photo-1511497584788-87676104235f?auto=format&fit=crop&w=1200&q=80
canonical_url:

description: "BioTrail AI is an offline, open-source AI assistant that helps developers step away from screens, complete sensory nature quests, and identify biodiversity without cell service."

This article is a submission for the Hacktoberfest Open-Source AI Challenge: Week 1 ("Touch Grass").

🌿 The Inspiration: What if AI pushed us away from screens?

As developers, we spend hours staring at monitors, reviewing pull requests, and debugging stacks. "Touch grass" started as an internet meme, but behind the joke lies a genuine need: human beings need regular sensory immersion in the real, physical world to stay healthy and creative.

However, existing nature apps come with two major flaws:

Cloud Lock-in & Cellular Blackouts: When you actually hike into a deep forest ravine or mountain trail, cell signal drops to zero. Cloud-based proprietary AI APIs fail completely.
Privacy & Telemetry: Why should a hiker's exact GPS location, photos of secluded local foraging spots, and personal audio reflections be shipped to advertising-funded big tech servers?

I built BioTrail AI to flip the script: an open-source, local-first field assistant powered by open-weight AI that works entirely offline.

🚀 What I Built

BioTrail AI is an open-source companion designed for wild trails and neighborhood parks:

📷 Botanical & Ecological Vision Identification: Point your camera at any wild plant, leaf, mushroom, or bark texture. The local vision pipeline identifies the specimen, classifies its taxonomic family, explains its ecological role, and shares naturalist fun facts.
🎙️ Screenless Audio Nature Journal: Walk with your phone tucked away. Using voice dictation and natural language extraction, the app listens to your spoken reflections and automatically tags sensory landmarks (such as #avifauna, #weather, #flora, #mindfulness).
🧭 Procedural "Touch Grass" Quests: Rather than passive scrolling, an open-source quest agent serves micro-adventures (e.g., "Locate an angiosperm with serrated leaf margins", "Perform a 10-minute acoustic forest bath", "Inspect an urban pollinator patch").
🏆 Offline XP & Badges: Gamifies real-world outdoor time. The app tracks your total outdoor minutes, logs species into a personal offline herbarium, and awards badges like Grass Toucher, Trail Pioneer, and Master Botanist.
📝 Markdown Field Guide Export: Generates clean, formatted Markdown logs of your biodiversity expeditions.

🧠 Why Open-Source AI was Integral to This Project

The Hacktoberfest challenge specifically asks why an open-source AI approach was essential. For BioTrail AI, open-source AI was not merely an option—it was the only viable architecture:

1. Zero-Cellular Resilience on Wilderness Trails

Nature doesn't have 5G towers. When you are deep inside an old-growth forest or national park, calling proprietary cloud APIs like OpenAI or Anthropic is impossible. Open-weight models (running locally via Ollama with quantized weights or through edge rule fallbacks) execute 100% on the device without an internet connection.

2. Radical Location Privacy

Outdoor exploration often involves sharing coordinates, foraging sites, and personal spoken reflections. By keeping all inference local and persisting observations in a private SQLite database (biotrail.db), user data never leaves their hardware.

3. Open Innovation & No Per-Query Toll Booths

Encouraging millions of young explorers and students to explore biology shouldn't incur API bills or rate limits. Open weights mean students and schools can run educational field tools at zero ongoing marginal cost.

🛠️ Tech Stack & Architecture

   ┌─────────────────────────────────────────────────────────┐
   │             Responsive Web Interface                    │
   │  (Tailwind CSS, Phosphor Icons, Web Speech Audio API)   │
   └────────────────────────────┬────────────────────────────┘
                                │ JSON API
   ┌────────────────────────────▼────────────────────────────┐
   │                 FastAPI Python Server                   │
   │        (Async REST endpoints, Local SQLite DB)          │
   └───────────────┬─────────────────────────┬───────────────┘
                   │                         │
   ┌───────────────▼──────────────┐  ┌───────▼───────────────┐
   │    Open-Source AI Engine     │  │   Offline Herbarium   │
   │ ├─ Ollama (llama3.2-vision)  │  │   (Local SQLite DB    │
   │ └─ Edge Taxonomy Knowledge   │  │    biotrail.db)       │
   └──────────────────────────────┘  └───────────────────────┘

Backend: FastAPI, Python 3.12, Uvicorn, Pydantic, HTTPX
AI Models: Ollama Open-Weights (llama3.2-vision, llama3.2) with deterministic edge fallback
Frontend: Vanilla JS, Tailwind CSS, Web Speech Recognition API
Persistence: Local SQLite database with zero cloud synchronization

💻 Code Spotlight: The Dual-Path Local Inference Engine

BioTrail AI seamlessly probes for a running Ollama daemon. If available, it invokes open-weight multimodal models (llama3.2-vision). If running on a lightweight laptop on a trail without a dedicated GPU, it gracefully delegates to the local edge taxonomy bank:

async def identify_flora_image(base64_img: str, user_notes: str = "") -> Dict[str, Any]:
    is_ollama_up = await check_ollama_available()

    if is_ollama_up:
        try:
            payload = {
                "model": "llama3.2-vision:latest",
                "prompt": "You are an expert field botanist AI. Identify this specimen and return JSON...",
                "images": [base64_img.split(",")[-1]],
                "stream": False,
                "format": "json"
            }
            async with httpx.AsyncClient(timeout=30.0) as client:
                res = await client.post("http://127.0.0.1:11434/api/generate", json=payload)
                if res.status_code == 200:
                    parsed = json.loads(res.json()["response"])
                    parsed["ai_engine"] = "Local Ollama (llama3.2-vision)"
                    return parsed
        except Exception:
            pass

    # Seamless Edge Fallback: zero dependencies, zero network required!
    return fallback_botanical_taxonomy(user_notes)

🏃 Running It Yourself

Prerequisites

Python 3.10+
(Optional) Ollama with llama3.2-vision installed

Quick Setup

# 1. Clone repository
git clone https://github.com/your-username/biotrail-ai.git
cd biotrail-ai

# 2. Install Python packages
cd backend
pip install -r requirements.txt

# 3. Start the server
python main.py

Now open http://localhost:8000 in your browser, strap on your walking shoes, and step outside!

🗺️ What's Next for BioTrail AI

[ ] Native PWA packaging with local WebAssembly Whisper model for offline speech recognition.
[ ] Direct GPX route map rendering with heatmaps of high-biodiversity microclimates.
[ ] Community mesh sync (peer-to-peer Wi-Fi Direct or Bluetooth sync between hikers on the same trail).

📦 Source Code & Repository

GitHub Repository: https://github.com/your-username/biotrail-ai
License: MIT

Thank you to DEV and the Hacktoberfest organizers for this challenge! Let's close our laptops, breathe some fresh air, and touch grass! 🌱