<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Ankush Chakraborty</title>
    <description>The latest articles on DEV Community by Ankush Chakraborty (@ankxiety).</description>
    <link>https://dev.to/ankxiety</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4154717%2F039b4a54-33c0-4c17-ad8e-8596992e1f02.png</url>
      <title>DEV Community: Ankush Chakraborty</title>
      <link>https://dev.to/ankxiety</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/ankxiety"/>
    <language>en</language>
    <item>
      <title>GhostWrite AI</title>
      <dc:creator>Ankush Chakraborty</dc:creator>
      <pubDate>Wed, 28 Oct 2026 10:14:22 +0000</pubDate>
      <link>https://dev.to/ankxiety/ghostwrite-ai-4pk2</link>
      <guid>https://dev.to/ankxiety/ghostwrite-ai-4pk2</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week4-2026-10-26"&gt;Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;What I Built&lt;/p&gt;

&lt;p&gt;GhostWrite AI is a spooky AI-powered horror story generator inspired by the theme "Ghost in the Machine."&lt;/p&gt;

&lt;p&gt;Users can enter a story idea, choose a haunting style, select the length and intensity, and generate an original horror story. The app features a dark, terminal-inspired interface and lets users download their finished stories.&lt;/p&gt;

&lt;p&gt;I wanted to create something atmospheric while keeping the underlying technology approachable and usable on an ordinary personal computer.&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;


&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/ANILION-CODE" rel="noopener noreferrer"&gt;
        ANILION-CODE
      &lt;/a&gt; / &lt;a href="https://github.com/ANILION-CODE/GhostWrite_AI" rel="noopener noreferrer"&gt;
        GhostWrite_AI
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      GhostWrite AI is a spooky AI-powered horror story generator inspired by the theme "Ghost in the Machine."  Users can enter a story idea, choose a haunting style, select the length and intensity, and generate an original horror story. The app features a dark, terminal-inspired interface and lets users download their finished stories.
    &lt;/h3&gt;
  &lt;/div&gt;
&lt;/div&gt;


&lt;p&gt;How I Built It&lt;/p&gt;

&lt;p&gt;I built GhostWrite AI with Python and Streamlit, using Ollama to communicate with an open-weight language model running locally.&lt;/p&gt;

&lt;p&gt;The application sends a story-generation prompt to the local Ollama API and displays the model's response in a haunted terminal-style interface. Users can customise their stories and download them as text files.&lt;/p&gt;

&lt;p&gt;Why Does Open Innovation Matter?&lt;/p&gt;

&lt;p&gt;I wanted the story-generation engine to run on my own computer rather than depend on a proprietary, hosted story-generation API.&lt;/p&gt;

&lt;p&gt;Using a locally running model means the application can generate stories without sending prompts to a third-party story-generation service. It also gives me the flexibility to experiment with other compatible models.&lt;/p&gt;

&lt;p&gt;Once the model and software have been installed, local inference can work without an internet connection. This makes the project less dependent on an external inference service and gives the user more control over their prompts.&lt;/p&gt;

&lt;p&gt;The project also helped me explore how open-weight AI can power a complete application rather than just serve as a chatbot.&lt;/p&gt;

&lt;p&gt;The Spirit of It&lt;/p&gt;

&lt;p&gt;GhostWrite AI treats the computer as a doorway into a fictional haunted world. A simple prompt becomes a story, and the familiar terminal becomes part of the atmosphere.&lt;/p&gt;

&lt;p&gt;The stories are generated fiction, but the experience celebrates the strange and creative possibilities of open AI.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>LazyDesk AI</title>
      <dc:creator>Ankush Chakraborty</dc:creator>
      <pubDate>Sat, 24 Oct 2026 13:59:26 +0000</pubDate>
      <link>https://dev.to/ankxiety/lazydesk-ai-2mbm</link>
      <guid>https://dev.to/ankxiety/lazydesk-ai-2mbm</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week3-2026-10-19"&gt;Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;What I Built&lt;/p&gt;

&lt;p&gt;LazyDesk AI is a lightweight AI-powered productivity assistant that transforms messy notes and brain dumps into a concise summary, a prioritized checklist, and a practical next-step plan.&lt;/p&gt;

&lt;p&gt;Instead of spending time organizing every task manually, users can paste their notes, choose a planning style, and let AI do the organizing.&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;


&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/ANILION-CODE" rel="noopener noreferrer"&gt;
        ANILION-CODE
      &lt;/a&gt; / &lt;a href="https://github.com/ANILION-CODE/LazyDesk_AI" rel="noopener noreferrer"&gt;
        LazyDesk_AI
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      LazyDesk AI is a lightweight AI-powered productivity assistant that transforms messy notes and brain dumps into a concise summary, a prioritized checklist, and a practical next-step plan.  Instead of spending time organizing every task manually, users can paste their notes, choose a planning style, and let AI do the organizing.
    &lt;/h3&gt;
  &lt;/div&gt;
&lt;/div&gt;


&lt;p&gt;How I Built It&lt;/p&gt;

&lt;p&gt;I built LazyDesk AI with Python and Streamlit. For the AI engine, it uses the open-weight openai/gpt-oss-120b model through Hugging Face Inference Providers.&lt;/p&gt;

&lt;p&gt;The application sends the user's notes to the model and displays its generated summary, prioritized tasks, and suggested plan. Users can also download the resulting plan as a text file.&lt;/p&gt;

&lt;p&gt;The interface is deployed as a Hugging Face Space using Docker.&lt;/p&gt;

&lt;p&gt;Why Does Open Innovation Matter?&lt;/p&gt;

&lt;p&gt;Open-weight AI makes it possible to build useful applications around models that developers can inspect, adapt, and integrate into their own projects.&lt;/p&gt;

&lt;p&gt;For LazyDesk AI, the model is a replaceable component rather than a closed, hardwired dependency. The application code is publicly available, and other developers can adapt the prompts, interface, and model choice.&lt;/p&gt;

&lt;p&gt;This demo uses hosted inference, so it is not fully offline or entirely local. Notes are sent to the inference service, and usage depends on the provider's current availability and terms.&lt;/p&gt;

&lt;p&gt;The Lazy Payoff&lt;/p&gt;

&lt;p&gt;LazyDesk AI aims to reduce the small but repetitive effort of turning scattered thoughts into an actionable plan.&lt;/p&gt;

&lt;p&gt;I tested the workflow with example task notes and used the generated checklist to demonstrate how the tool can turn an unstructured brain dump into a clearer starting point.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>NoteWise_AI</title>
      <dc:creator>Ankush Chakraborty</dc:creator>
      <pubDate>Fri, 16 Oct 2026 13:10:48 +0000</pubDate>
      <link>https://dev.to/ankxiety/notewiseai-g14</link>
      <guid>https://dev.to/ankxiety/notewiseai-g14</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week2-2026-10-12"&gt;Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;I built NoteWise AI, a local-first AI study assistant designed to help students study more efficiently using their own notes.&lt;/p&gt;

&lt;p&gt;Students often have to go through lengthy PDFs and lecture notes to prepare for exams. NoteWise AI aims to make this process easier by allowing students to upload their notes and use AI to:&lt;/p&gt;

&lt;p&gt;Generate short summaries and exam-oriented revision notes.&lt;/p&gt;

&lt;p&gt;Ask questions about their study material.&lt;/p&gt;

&lt;p&gt;Generate practice quizzes and answer keys.&lt;/p&gt;

&lt;p&gt;Download generated summaries and quizzes for revision.&lt;/p&gt;

&lt;p&gt;The initial version supports text-based PDF and TXT files.&lt;/p&gt;

&lt;p&gt;The project is intended for students, independent learners, and tutors. In the future, it could become a low-cost study tool for students who want useful AI features without paying for an expensive subscription.&amp;gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;


&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/ANILION-CODE" rel="noopener noreferrer"&gt;
        ANILION-CODE
      &lt;/a&gt; / &lt;a href="https://github.com/ANILION-CODE/NoteWise_AI" rel="noopener noreferrer"&gt;
        NoteWise_AI
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      A local first study assistant that uses an open*weight language model through Ollama to summarize notes, answer questions from uploaded material, and generate practice quizzes.
    &lt;/h3&gt;
  &lt;/div&gt;
  &lt;div class="ltag-github-body"&gt;
    
&lt;div id="readme" class="md"&gt;&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;NoteWise AI&lt;/h1&gt;
&lt;/div&gt;
&lt;p&gt;A local first study assistant that uses an open*weight language model through Ollama to summarize notes, answer questions from uploaded material, and generate practice quizzes.&lt;/p&gt;
&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;Features&lt;/h1&gt;
&lt;/div&gt;
&lt;ul&gt;
&lt;li&gt;Upload selectable*text PDF or TXT notes&lt;/li&gt;
&lt;li&gt;Generate short, detailed, or exam revision summaries&lt;/li&gt;
&lt;li&gt;Ask questions using a lightweight keyword*based context selector&lt;/li&gt;
&lt;li&gt;Generate practice quizzes and download results as Markdown&lt;/li&gt;
&lt;li&gt;Run inference through Ollama on your own computer&lt;/li&gt;
&lt;/ul&gt;
&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;Requirements&lt;/h1&gt;
&lt;/div&gt;
&lt;ul&gt;
&lt;li&gt;Windows 10/11&lt;/li&gt;
&lt;li&gt;Python 3.10 or newer&lt;/li&gt;
&lt;li&gt;Ollama for Windows&lt;/li&gt;
&lt;li&gt;Enough RAM to run the model you choose&lt;/li&gt;
&lt;/ul&gt;
&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;Run locally&lt;/h1&gt;

&lt;/div&gt;
&lt;ol&gt;
&lt;li&gt;
&lt;p&gt;Install Python from &lt;a href="https://www.python.org/downloads/" rel="nofollow noopener noreferrer"&gt;https://www.python.org/downloads/&lt;/a&gt; if you do not already have it.&lt;/p&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;p&gt;Install Ollama from &lt;a href="https://ollama.com/download" rel="nofollow noopener noreferrer"&gt;https://ollama.com/download&lt;/a&gt;.&lt;/p&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;p&gt;Open Command Prompt or PowerShell and download the default model:&lt;/p&gt;
&lt;div class="highlight highlight-source-powershell notranslate position-relative overflow-auto js-code-highlight"&gt;
&lt;pre&gt;ollama run llama3.&lt;span class="pl-c1"&gt;2&lt;/span&gt;:3b&lt;/pre&gt;

&lt;/div&gt;
&lt;p&gt;Wait for the model to download. When the model prompt appears, you can type &lt;code&gt;/bye&lt;/code&gt; to exit. Leave Ollama running in the background.&lt;/p&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;p&gt;Extract this repository, open a terminal in the &lt;code&gt;notewise_ai&lt;/code&gt; folder, and…&lt;/p&gt;
&lt;/li&gt;
&lt;/ol&gt;&lt;/div&gt;
  &lt;/div&gt;
  &lt;div class="gh-btn-container"&gt;&lt;a class="gh-btn" href="https://github.com/ANILION-CODE/NoteWise_AI" rel="noopener noreferrer"&gt;View on GitHub&lt;/a&gt;&lt;/div&gt;
&lt;/div&gt;


&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;I built NoteWise AI using Python and Streamlit, with Ollama providing access to a locally running open-weight language model.&lt;/p&gt;

&lt;p&gt;The main technologies are:&lt;/p&gt;

&lt;p&gt;Python: Application logic and document processing.&lt;/p&gt;

&lt;p&gt;Streamlit: Browser-based user interface.&lt;/p&gt;

&lt;p&gt;Ollama: Runs the language model locally.&lt;/p&gt;

&lt;p&gt;Llama 3.2 3B: The default model configuration for generating summaries, answers, and quizzes.&lt;/p&gt;

&lt;p&gt;pypdf: Extracts text from PDF documents.&lt;/p&gt;

&lt;p&gt;Lightweight keyword retrieval: Selects relevant text from notes when answering questions.&lt;/p&gt;

&lt;p&gt;The application sends prompts to the local Ollama service instead of relying on a hosted AI API. The initial version uses a simple retrieval approach, which I plan to improve with embeddings, vector search, and source references.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Open innovation is important to NoteWise AI because students should have more choices in how they use AI tools and where their data is processed.&lt;/p&gt;

&lt;p&gt;Using an open-weight model through Ollama makes local inference possible. In the intended local setup, study documents and prompts can remain on the user's computer rather than being sent to a third-party AI API.&lt;/p&gt;

&lt;p&gt;It also gives the project flexibility. Developers can inspect and modify the code, experiment with compatible models, and improve retrieval or prompting for different subjects.&lt;/p&gt;

&lt;p&gt;Running a model locally can reduce recurring per-request API charges, although hardware, electricity, and setup still have costs. This could make NoteWise AI useful for students who want to avoid depending on paid AI subscriptions.&lt;/p&gt;

&lt;p&gt;The goal is to build a tool that users can understand, adapt, and improve instead of locking the entire application to one proprietary AI provider.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>GreenBuddy</title>
      <dc:creator>Ankush Chakraborty</dc:creator>
      <pubDate>Sun, 11 Oct 2026 09:53:40 +0000</pubDate>
      <link>https://dev.to/ankxiety/greenbuddy-2533</link>
      <guid>https://dev.to/ankxiety/greenbuddy-2533</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week1-2026-10-05"&gt;Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;GreenBuddy is an AI-powered gardening assistant designed to help beginners turn gardening questions into real-world action.&lt;/p&gt;

&lt;p&gt;Users can enter their city, available growing space, sunlight conditions, and the plants they want to grow. GreenBuddy generates a personalized gardening plan with practical care tips and an outdoor mission to complete.&lt;/p&gt;

&lt;p&gt;The idea is simple: spend less time looking at a screen and more time growing something in the real world&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;


&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/ANILION-CODE" rel="noopener noreferrer"&gt;
        ANILION-CODE
      &lt;/a&gt; / &lt;a href="https://github.com/ANILION-CODE/GreenBuddy" rel="noopener noreferrer"&gt;
        GreenBuddy
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      GreenBuddy is an AI-powered gardening assistant designed to help beginners turn gardening questions into real-world action.  Users can enter their city, available growing space, sunlight conditions, and the plants they want to grow. GreenBuddy generates a personalized gardening plan with practical care tips and an outdoor mission to complete. 
    &lt;/h3&gt;
  &lt;/div&gt;
&lt;/div&gt;


&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;GreenBuddy is built with Python and Streamlit, with Hugging Face Inference powering its AI gardening assistant.&lt;/p&gt;

&lt;p&gt;The prototype collects basic information about a user's growing environment and sends it to an open-weight language model to generate personalized gardening guidance.&lt;/p&gt;

&lt;p&gt;The application is designed to be beginner-friendly, with a simple interface and practical recommendations instead of an unnecessarily complicated setup.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Open innovation makes it possible to experiment with AI without depending entirely on proprietary models or building expensive infrastructure.&lt;/p&gt;

&lt;p&gt;By using an open-weight model through hosted inference, GreenBuddy can explore model choice and customization while keeping the application code accessible to other developers.&lt;/p&gt;

&lt;p&gt;The project also leaves room for future improvements, including alternative models, local inference, and a gardening journal that users control.&lt;/p&gt;

&lt;p&gt;My goal is to make AI useful beyond the screen: helping people learn about plants, spend time outdoors, and build a gardening habit.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>first post</title>
      <dc:creator>Ankush Chakraborty</dc:creator>
      <pubDate>Thu, 01 Oct 2026 12:21:52 +0000</pubDate>
      <link>https://dev.to/ankxiety/first-post-2471</link>
      <guid>https://dev.to/ankxiety/first-post-2471</guid>
      <description>&lt;p&gt;this must be done&lt;/p&gt;

</description>
    </item>
  </channel>
</rss>
