The latest articles on DEV Community by Rittik Chakraborty (@devrittik). https://dev.to/devrittik 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%2F4097784%2F0f96e453-fda6-475c-ab6e-9b7c24d34c49.png https://dev.to/devrittik en Rittik Chakraborty Sun, 01 Nov 2026 05:56:12 +0000 https://dev.to/devrittik/link-seance-summoning-the-ghosts-of-forgotten-websites-with-open-weight-ai-1e0c https://dev.to/devrittik/link-seance-summoning-the-ghosts-of-forgotten-websites-with-open-weight-ai-1e0c <p><em>This is a submission for the <a href="https://dev.to/challenges/hacktoberfest-week4-2026-10-26">Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine</a></em></p> <h2> What I Built </h2> <p><strong>Link Séance — Summon the ghosts of forgotten web pages.</strong></p> <p>Ever wondered what happened to an old website? Maybe a blog you used to visit, an abandoned personal website, or a page that disappeared from the internet years ago?</p> <p>Link Séance is a small web application that brings those forgotten corners of the internet back to life.</p> <p>Give it an old or dead URL, and it:</p> <ul> <li>Searches the Internet Archive's Wayback Machine for the closest available snapshot.</li> <li>Fetches the archived webpage and extracts its readable content.</li> <li>Uses an open-weight LLM to reconstruct the story behind the page.</li> <li>Explains what the website was about, who it was intended for, and potentially suggests a modern alternative when the model has enough confidence.</li> <li>Offers an optional <strong>Ghost Mode</strong>, which narrates the result in a gentle, spooky tone.</li> </ul> <p>I interpreted the Ghost in the Machine theme quite literally: what if the ghosts haunting the internet were not supernatural, but the forgotten websites, communities, and digital memories we left behind?</p> <p>The idea was inspired by the thought that every dead link has a story worth discovering.</p> <h2> Demo </h2> <p>Try it here:</p> <p>🔗 <a href="https://link-seance.vercel.app" rel="noopener noreferrer">https://link-seance.vercel.app</a></p> <h2> Code </h2> <p>The project is open source and available on GitHub:</p> <p>🔗 <a href="https://github.com/devrittik/link-seance" rel="noopener noreferrer">https://github.com/devrittik/link-seance</a></p> <p>Built with a modular architecture so the application can be extended without tightly coupling the UI, archive retrieval, and AI inference.</p> <h2> How I Built It </h2> <p>I wanted to keep the project lightweight and avoid depending on a proprietary AI API.</p> <p><strong>Tech stack:</strong></p> <ul> <li>Next.js 16 and App Router</li> <li>TypeScript</li> <li>Tailwind CSS v4</li> <li>shadcn/ui</li> <li>Cheerio for HTML content extraction</li> <li>Internet Archive's Wayback Machine</li> <li>llama.cpp for local LLM inference</li> </ul> <p><strong>The AI integration was designed to be provider-independent.</strong></p> <p>The application communicates with an OpenAI-compatible chat completions endpoint. This means I can run an open-weight model locally through llama.cpp or switch to a compatible hosted inference provider simply by changing environment variables.</p> <p>The AI configuration uses:</p> <ul> <li><code>LLM_BASE_URL</code></li> <li><code>LLM_MODEL</code></li> <li><code>LLM_API_KEY</code></li> </ul> <p>No application code changes are needed when switching between compatible inference providers.</p> <p>The main workflow looks like this:</p> <ol> <li>User submits an old or dead URL.</li> <li>The backend searches the Wayback Machine for an archived snapshot.</li> <li>The archived HTML is fetched and cleaned using Cheerio.</li> <li>Extracted content is sent to the configured LLM.</li> <li>The model generates a contextual explanation of the forgotten website.</li> <li>The result is presented with the original snapshot date and an optional spooky narration mode.</li> </ol> <p>I also added URL validation, distinct error responses, and request timeouts to make the experience more reliable.</p> <h2> Why Does Open Innovation Matter? </h2> <p>For this project, open innovation isn't just about using open-source tools. It's about keeping control over the entire AI pipeline.</p> <p>Using an open-weight model with llama.cpp means the application can run without depending on a proprietary AI API, and potentially without sending archived webpage content to a third-party inference provider.</p> <p>It also makes experimentation easier.</p> <p>Developers can:</p> <ul> <li>Choose and experiment with different open-weight models.</li> <li>Run inference locally on their own hardware.</li> <li>Switch between local and hosted inference.</li> <li>Modify the prompts and model configuration.</li> <li>Build on the project without being locked into a single AI provider.</li> </ul> <p>I wanted Link Séance to be something that developers could actually run, inspect, and adapt.</p> <h2> My Agent Session </h2> <p>Coming soon!</p> <h2> The Spirit of It </h2> <p>The internet has a strange kind of memory.</p> <p>A website can disappear, its owner can move on, and its community can scatter. But sometimes a snapshot survives somewhere in the archives.</p> <p>I like the idea that a forgotten URL can become a doorway into a different time.</p> <p>Link Séance is my attempt to make exploring that digital past a little more interactive — and perhaps a little spooky.</p> <p><strong>Not every ghost is dead. Some are just waiting in the archive.</strong> 👻</p> <h2> Prize Categories </h2> <p>Open-Source AI Challenge — Week 4: Ghost in the Machine.</p> devchallenge hf26challenge Rittik Chakraborty Mon, 26 Oct 2026 06:58:23 +0000 https://dev.to/devrittik/returnrakshak-never-miss-a-return-window-again-4i7k https://dev.to/devrittik/returnrakshak-never-miss-a-return-window-again-4i7k <p><em>This is a submission for the <a href="https://dev.to/challenges/hacktoberfest-week3-2026-10-19">Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady</a></em></p> <h1> ReturnRakshak: Never Miss a Return Window Again 🛡️ </h1> <h2> What I Built </h2> <p>Ever bought something online and forgotten when the return window closes? Or struggled to find a warranty invoice months later?</p> <p>Meet <strong>ReturnRakshak</strong> — a privacy-first AI invoice assistant designed to reduce the mental overhead of tracking product returns and warranties.</p> <p>The idea is simple: upload an invoice, extract the important details with AI, review the results, and keep track of important deadlines without manually entering everything.</p> <p>The goal is to make everyday purchase management a little less stressful and let people spend less time organizing documents.</p> <h2> Demo </h2> <div class="crayons-card c-embed text-styles text-styles--secondary"> <div class="c-embed__content"> <div class="c-embed__body"> <h2 class="fs-xl lh-tight"> <a href="https://drive.google.com/file/d/1ZNnzyoPGFdGvrNkpaqtYl-CubMsboENM/view?usp=sharing" rel="noopener noreferrer" class="c-link"> Project 2.mp4 - Google Drive </a> </h2> <div class="color-secondary fs-s flex items-center"> <img alt="favicon" class="c-embed__favicon m-0 mr-2 radius-0" src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fssl.gstatic.com%2Fdocs%2Fdoclist%2Fimages%2Fdrive_favicon_2026_32dp.png" width="32" height="32"> drive.google.com </div> </div> </div> </div> <h2> Code </h2> <p>🔗 <a href="https://github.com/devrittik/return-rakshak" rel="noopener noreferrer">GitHub Repository — devrittik/return-rakshak</a></p> <h2> How I Built It </h2> <p>ReturnRakshak is designed around a modular, privacy-first architecture.</p> <p>The planned technology stack includes:</p> <ul> <li> <strong>Next.js and TypeScript</strong> for the application.</li> <li> <strong>Tailwind CSS</strong> for a responsive, installable PWA experience.</li> <li> <strong>PDF.js and Tesseract.js</strong> for extracting text from digital and scanned invoices.</li> <li> <strong>Open-weight language models</strong> for converting extracted text into structured invoice information.</li> <li> <strong>SQLite/Turso with Drizzle ORM</strong> for storing structured product, policy, and reminder data.</li> </ul> <p>A key design decision is to avoid permanently storing original invoices. Documents are processed temporarily, and only the necessary structured information is retained after user confirmation.</p> <p>AI handles information extraction; deterministic application logic handles deadline calculations. Missing or uncertain policy details should be reviewed rather than guessed.</p> <h2> Why Does Open Innovation Matter? </h2> <p>Privacy-sensitive documents shouldn't require users to depend entirely on a proprietary AI service.</p> <p>Open-weight models and local inference offer the possibility of processing invoice data without sending it to a third-party AI API. They also give developers more control over model selection, deployment, and experimentation.</p> <p>Open innovation makes it possible to build useful everyday tools while keeping implementation choices transparent and adaptable.</p> <h2> My Agent Session </h2> <p>Built with AI-assisted development.</p> <p>Add your DevRelay session link here if you recorded and saved one. Otherwise, remove this section before publishing.</p> <h2> The Lazy Payoff </h2> <p>The goal is to replace repetitive manual entry with a simple workflow: upload, review, confirm, and track.</p> <p>Instead of repeatedly searching through invoices or trying to remember expiry dates, users can keep important purchase information organized in one place.</p> <p>The time saved will depend on the number of purchases and the quality of invoice extraction. I plan to evaluate this using sample invoices rather than inventing a time-saving figure.</p> <h2> Prize Categories </h2> <h1> hf26challenge #opensource #AI #NextJS #Hacktoberfest </h1> devchallenge hf26challenge Rittik Chakraborty Mon, 19 Oct 2026 06:53:22 +0000 https://dev.to/devrittik/estiforge-building-an-open-source-ai-engineering-estimation-assistant-20n3 https://dev.to/devrittik/estiforge-building-an-open-source-ai-engineering-estimation-assistant-20n3 <p><em>This is a submission for the <a href="https://dev.to/challenges/hacktoberfest-week2-2026-10-12">Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle</a></em></p> <h1> EstiForge: Open-Source AI for Engineering Estimation </h1> <h2> What I Built </h2> <p>EstiForge is an AI-assisted engineering estimation platform designed for independent engineers and small contractors.</p> <p>The goal is to simplify the process of turning project requirements into structured estimates, Bills of Quantities (BOQs), and professional quotations.</p> <p>Instead of manually preparing every estimate from scratch, users can describe their requirements, review the extracted inputs, calculate quantities, manage material and labour rates, and generate quotations.</p> <p>I'm starting with electrical estimation, with a modular architecture that can expand to civil engineering and other trades.</p> <h2> Demo </h2> <p><strong>Live Demo:</strong> <a href="https://estiforge.vercel.app" rel="noopener noreferrer">EstiForge — Live Application</a></p> <h2> Code </h2> <p><strong>GitHub Repository:</strong> <a href="https://github.com/devrittik/estiforge" rel="noopener noreferrer">EstiForge</a></p> <h2> How I Built It </h2> <p>EstiForge is designed around open-source AI and a modular architecture.</p> <ul> <li> <strong>Next.js + TypeScript:</strong> Application framework and type-safe business logic.</li> <li> <strong>Tailwind CSS + shadcn/ui:</strong> Responsive interface and reusable components.</li> <li> <strong>llama.cpp:</strong> Local inference using compatible open-weight language models.</li> <li> <strong>MongoDB:</strong> Persistent project, estimate, and quotation data.</li> <li> <strong>Formula Registry:</strong> A modular approach to engineering calculations, with deterministic code for supported formulas.</li> <li> <strong>Voice Input:</strong> Planned local speech recognition using Whisper through whisper.cpp.</li> </ul> <p>The LLM helps interpret natural-language requirements and extract structured inputs. Engineering calculations and final pricing are handled separately so that model-generated text cannot silently determine the final estimate.</p> <h2> Why Does Open Innovation Matter? </h2> <p>Engineering projects can contain sensitive customer information, project specifications, and commercial rates.</p> <p>Open-weight models and local inference offer the possibility of processing that information without sending every request to a third-party AI provider. They also give developers the freedom to change models, inspect their behaviour, and adapt the application to specific workflows.</p> <p>Using open-source tools makes it possible to build a flexible estimation platform without depending entirely on proprietary AI APIs or their pricing.</p> <p>My long-term goal is to make practical AI-assisted engineering tools more accessible to independent professionals and small businesses.</p> <h2> Prize Categories </h2> <ul> <li>Open-Source AI</li> <li>Side Hustle</li> </ul> devchallenge hf26challenge Rittik Chakraborty Sun, 11 Oct 2026 14:26:27 +0000 https://dev.to/devrittik/i-built-an-ai-that-teaches-you-to-notice-the-world-not-stare-at-a-screen-10bp https://dev.to/devrittik/i-built-an-ai-that-teaches-you-to-notice-the-world-not-stare-at-a-screen-10bp <p>This is a submission for the <a href="https://dev.to/challenges/hacktoberfest-week1-2026-10-05">Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass</a>.</p> <h2> What I Built </h2> <p><strong>Observation Coach — A little less screen. A little more world.</strong></p> <p>Most AI applications try to keep you engaged with your screen. I wanted to build something that does the opposite: an AI coach that helps you put your phone away and pay closer attention to the world around you.</p> <p><a href="https://observation-coach.vercel.app" rel="noopener noreferrer">Observation Coach — Live Preview</a></p> <p>The idea is simple: <strong>AI should teach you how to observe, rather than observe everything for you.</strong></p> <p>Here's how it works:</p> <ol> <li> <strong>Choose an exercise:</strong> Pick a short observation session lasting 5, 10, 20, or 30 minutes.</li> <li> <strong>Step away:</strong> The app invites you to put your phone away and explore your surroundings.</li> <li> <strong>Notice something real:</strong> Return and describe a detail you actually observed.</li> <li> <strong>Reflect with AI:</strong> The coach asks a thoughtful follow-up question based on your description.</li> <li> <strong>Build the skill:</strong> Receive an adaptive next exercise designed to help you develop your observation skills over time.</li> </ol> <p>Instead of identifying objects through a camera, generating endless challenges, or turning everything into a competition, Observation Coach focuses on attention, description, comparison, change, inference, and deeper observation.</p> <p>It's designed for anyone who wants to reconnect with their surroundings, become more mindful of everyday details, or simply spend less time staring at a screen.</p> <p>The goal isn't to make people spend more time with AI. It's to make them need AI less.</p> <h2> Demo </h2> <p><strong>Product Walkthrough:</strong><br><br> <a href="https://youtu.be/KArn5bRtpeI" rel="noopener noreferrer">https://youtu.be/KArn5bRtpeI</a></p> <p><strong>Code Walkthrough:</strong><br><br> <a href="https://youtu.be/vQHZ7amHMTo" rel="noopener noreferrer">https://youtu.be/vQHZ7amHMTo</a></p> <p><strong>Try the Live Preview:</strong><br><br> <a href="https://observation-coach.vercel.app" rel="noopener noreferrer">https://observation-coach.vercel.app</a></p> <h2> Code </h2> <p>The project is open source:</p> <p><a href="https://github.com/devrittik/observation-coach" rel="noopener noreferrer">https://github.com/devrittik/observation-coach</a></p> <p>Built with Next.js, TypeScript, Tailwind CSS, and MongoDB.</p> <h2> How I Built It </h2> <p>I built Observation Coach with a focus on modular architecture, privacy, and the flexibility to use open-weight models locally or through a hosted inference provider.</p> <h3> The technology stack </h3> <ul> <li> <strong>Frontend and application:</strong> Next.js App Router, TypeScript, and Tailwind CSS.</li> <li> <strong>Database:</strong> MongoDB for user accounts, sessions, saved observations, and progress.</li> <li> <strong>AI integration:</strong> A provider abstraction that supports local inference and hosted OpenAI-compatible APIs.</li> <li> <strong>Local inference:</strong> Qwen 3B Instruct in GGUF format, downloaded from Hugging Face and run through llama.cpp.</li> <li> <strong>Hosted inference:</strong> Qwen 27B Instruct through Groq.</li> <li> <strong>Validation:</strong> Structured model responses are validated before being used by the application.</li> </ul> <h3> Why the AI is a coach, not an answer machine </h3> <p>The interesting part of this project isn't simply generating a response to a user's observation.</p> <p>The coach needs to meet the user where they are. If someone notices only a basic detail, the AI should help them explore it further rather than immediately provide an elaborate interpretation.</p> <p>The application uses structured prompts and validated responses to support short reflections, follow-up questions, and adaptive exercises. A separate progression system helps determine what the user can practise next.</p> <h3> Local AI and hosted AI </h3> <p>I wanted the project to be adaptable rather than permanently tied to one AI provider.</p> <p>The local setup uses an instruction-tuned Qwen model in GGUF format through llama.cpp. For hosted inference, the application supports an OpenAI-compatible API, with Qwen 27B Instruct through Groq as my intended hosted configuration.</p> <p>This separation means the application can support different inference environments without requiring a complete rewrite of its AI integration.</p> <p>The project also includes a guided demo mode, so people can explore the observation workflow without requiring a working AI API key.</p> <h2> Why Does Open Innovation Matter? </h2> <p>For Observation Coach, open innovation is not just a way to reduce API costs. It changes where and how the application can be used.</p> <h3> 1. AI that can work offline </h3> <p>With a local application, a local database, and a local open-weight model, the experience can work without an internet connection once everything is installed and running.</p> <p>That makes the concept useful even in places with limited or unreliable connectivity. It also means the user doesn't have to depend on a remote AI service just to practise observing their surroundings.</p> <h3> 2. Privacy and control over personal observations </h3> <p>An observation can be personal. People might write about their surroundings, emotions, memories, or everyday experiences.</p> <p>Running inference locally allows those observations to remain on the user's machine rather than being sent to an external AI provider. This is an important option for a product built around personal reflection.</p> <p>The hosted version is a practical alternative, but local inference gives users more control over their data.</p> <h3> 3. Freedom to experiment and switch models </h3> <p>Open-weight models let developers experiment with different model sizes, prompts, and inference setups.</p> <p>A smaller model can be useful for resource-constrained machines, while a larger hosted model can offer another option when local hardware is insufficient. This flexibility is especially valuable when building a project that should remain accessible to people with different hardware and budgets.</p> <h3> 4. Technology that encourages people to disconnect </h3> <p>The most important design decision is that the AI is not supposed to become another reason to stay online.</p> <p>Observation Coach deliberately keeps the interaction short. It gives the user an exercise, steps aside, and waits for them to return with something they actually noticed.</p> <p>For me, that's the real point of the Touch Grass challenge: using AI to help people reconnect with the world, rather than creating another experience that competes for their attention.</p> <h2> My Agent Session </h2> <p>I didn't use DevRelay for this project, so I don't have a DevRelay session to share.</p> <p>The code and implementation details are available in the <a href="https://github.com/devrittik/observation-coach" rel="noopener noreferrer">GitHub repository</a>, and the <a href="https://youtu.be/vQHZ7amHMTo" rel="noopener noreferrer">code walkthrough</a> provides another way to explore the project.</p> <h2> Prize Categories </h2> <ul> <li>Best Use of MongoDB Atlas</li> </ul> <h2> Submission Type </h2> <p>Individual submission.</p> <h2> Portfolio </h2> <h2> <a href="https://devrittik.com" rel="noopener noreferrer">https://devrittik.com</a> </h2> <p>Built for Hacktoberfest 2026 — Open-Source AI Challenge, Week 1: Touch Grass.</p> <h1> devchallenge #hf26challenge #hacktoberfest #opensource #ai #nextjs #mongodb #llama.cpp </h1> devchallenge hf26challenge hacktoberfest ai Rittik Chakraborty Mon, 05 Oct 2026 06:57:22 +0000 https://dev.to/devrittik/dukaandaar-ai-an-open-source-ai-assistant-for-small-businesses-114a https://dev.to/devrittik/dukaandaar-ai-an-open-source-ai-assistant-for-small-businesses-114a <p><em>This is a submission for the <a href="https://dev.to/challenges/hacktoberfest-weekend-2026-10-01">Hacktoberfest Weekend Challenge: Build for a Friend</a></em></p> <h2> What I Built </h2> <p>I built <strong>Dukaandaar AI</strong> for my friend <strong>Suman Mandal</strong>, who works as an accountant handling inventory-related operations.</p> <p>The idea started from seeing how much day-to-day business information can involve repetitive entries, calculations, inventory tracking, sales, purchases, and expenses.</p> <p>I wanted to build something that could make these tasks easier and more natural to interact with.</p> <p>Dukaandaar AI is an AI-powered business management application for small shops and businesses. It provides features for:</p> <ul> <li>📦 Inventory management</li> <li>🛒 Sales</li> <li>📥 Purchases</li> <li>💰 Expenses</li> <li>📊 Business reports</li> <li>📉 Losses and adjustments</li> <li>🤖 Natural-language interaction with business data</li> <li>🎙️ Voice-based interaction</li> </ul> <p>For example, instead of manually searching through the application, a user can ask:</p> <blockquote> <p>"How many units of this product are left?"</p> </blockquote> <p>or:</p> <blockquote> <p>"How much did I spend this month?"</p> </blockquote> <p>The project started as something I wanted to build for one real person, and then I decided to stretch the idea into a more complete tool that could potentially be useful for other small businesses as well.</p> <h2> Demo </h2> <div class="crayons-card c-embed text-styles text-styles--secondary"> <div class="c-embed__content"> <div class="c-embed__body flex items-center justify-between"> <a href="https://drive.google.com/drive/folders/1TAvQgogzpOgHoXlvOtDkIH-pr9HDBwGm?usp=drive_link" rel="noopener noreferrer" class="c-link fw-bold flex items-center"> <span class="mr-2">drive.google.com</span> </a> </div> </div> </div> <h2> Code </h2> <p><strong>Repository:</strong> <a href="https://github.com/devrittik/dukaandaar-ai" rel="noopener noreferrer">https://github.com/devrittik/dukaandaar-ai</a></p> <h2> Live </h2> <p><strong>URL:</strong> <a href="https://dukaandaar-ai.vercel.app" rel="noopener noreferrer">https://dukaandaar-ai.vercel.app</a></p> <h2> How I Built It </h2> <p>Dukaandaar AI is built with:</p> <ul> <li>React</li> <li>Node.js</li> <li>Express</li> <li>MongoDB</li> <li>REST APIs</li> <li>AI/LLM integrations</li> <li>Local AI components</li> <li>Speech-to-text / voice interaction</li> <li>Authentication and application logic</li> </ul> <p>A major design goal was to support <strong>both local/private and hosted deployments</strong>.</p> <p>For users who want maximum privacy, the application can be run locally with local AI components, allowing business data to remain on their own machine.</p> <p>For users who prefer convenience and remote access, the application can also be deployed as a hosted application.</p> <p>The AI is not intended to be just a chatbot sitting beside the application. The goal is to make AI an actual interface for interacting with structured business information.</p> <h2> Why Does Open Innovation Matter? </h2> <p>For a small business application, privacy can be extremely important because inventory, sales, purchases, and financial information can be sensitive.</p> <p>Open-source AI and local inference make it possible to build an AI-powered application where users can choose to keep their data and inference local instead of being completely dependent on a closed third-party API.</p> <p>It also gives developers more control over the AI layer. Models and inference approaches can be changed or upgraded without having to redesign the entire application around a single proprietary provider.</p> <p>For Dukaandaar AI, that flexibility is especially useful because the project supports both <strong>fully private/local usage</strong> and <strong>hosted usage</strong>.</p> <h2> My Agent Session </h2> <p>This project was <strong>not built using DevRelay</strong>, so I don't have an agent session to include.</p> <h2> Prize Categories </h2> <p>Best Use of ElevenLabs</p> <h2> Portfolio </h2> <p><a href="https://devrittik.com" rel="noopener noreferrer">https://devrittik.com</a></p> <h1> devchallenge #weekendchallenge #hf26challenge </h1> devchallenge weekendchallenge hf26challenge