The latest articles on DEV Community by Jungudo Muhammad Tukur (@hafeezmt). https://dev.to/hafeezmt 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%2F4184616%2F818de9be-3e7d-451c-b826-b5d3f1525053.jpg https://dev.to/hafeezmt en Jungudo Muhammad Tukur Thu, 15 Oct 2026 15:07:55 +0000 https://dev.to/hafeezmt/how-i-built-listingwriter-an-offline-ai-assistant-for-nigerian-real-estate-agents-165j https://dev.to/hafeezmt/how-i-built-listingwriter-an-offline-ai-assistant-for-nigerian-real-estate-agents-165j <h1> How I Built ListingWriter: An Offline AI Assistant for Nigerian Real Estate Agents </h1> <h2> What I Built </h2> <p>I developed <strong>ListingWriter</strong>, an open-source AI application designed to help Nigerian real estate agents create professional property advertisements more efficiently.</p> <p>ListingWriter generates three types of marketing content from a single property description:</p> <ul> <li>Professional real estate listings</li> <li>WhatsApp marketing messages</li> <li>Instagram captions with hashtags</li> </ul> <p>The application supports English, Nigerian Pidgin, and Hausa (beta).</p> <p>The goal is to make AI-powered marketing accessible to small real estate businesses, including agents who cannot afford expensive AI subscriptions.</p> <h2> Why I Built It </h2> <p>In Nigeria, real estate agents frequently advertise properties through WhatsApp, Instagram, and other digital platforms.</p> <p>Writing professional advertisements repeatedly takes time, especially when marketing several properties each day.</p> <p>I wanted to explore whether open-source AI could solve this problem while remaining affordable and privacy-focused.</p> <h2> How It Works </h2> <p>I developed ListingWriter using:</p> <ul> <li> <strong>Python:</strong> Application development</li> <li> <strong>Ollama:</strong> Local AI model execution</li> <li> <strong>Gemma 3 (1B):</strong> AI content generation</li> <li> <strong>Gradio:</strong> Interactive web interface</li> </ul> <p>Users enter information such as property type, location, price, bedrooms, bathrooms, and additional features.</p> <p>They can then generate marketing content based on those details.</p> <p>Because the AI runs locally, users do not need a paid cloud AI API for every generation.</p> <h2> Technical Challenges </h2> <p>One challenge was balancing model performance with the limited computing resources available on ordinary laptops.</p> <p>I selected the lightweight Gemma 3 1B model to make the application practical for computers without powerful graphics processors.</p> <p>Another challenge is language quality. Nigerian Pidgin and Hausa outputs can sometimes contain errors, so generated content should be reviewed before publication.</p> <h2> What I Learned </h2> <p>Building ListingWriter helped me improve my understanding of:</p> <ul> <li>Integrating open-weight AI models into practical applications</li> <li>Running AI locally using Ollama</li> <li>Designing user-friendly interfaces</li> <li>Developing technology for Nigerian business needs</li> <li>Working with multilingual AI generation</li> </ul> <h2> Project Demonstration </h2> <p>Watch my project demonstration:</p> <p> <iframe src="https://www.youtube.com/embed/UOiOf1euKgE" width="710" height="399"> </iframe> </p> <h2> GitHub Repository </h2> <p>The project source code is available here:</p> <p><a href="https://github.com/hafeezmt/listingwriter" rel="noopener noreferrer">https://github.com/hafeezmt/listingwriter</a></p> <h2> Future Improvements </h2> <p>My next development goals include:</p> <ul> <li>Better Hausa and Yoruba support</li> <li>Saving generated property advertisements</li> <li>Exporting marketing content as images</li> <li>Improving content consistency</li> <li>Exploring deployment options for real estate agencies</li> </ul> <h2> Final Thoughts </h2> <p>ListingWriter demonstrates how lightweight, open-weight AI models can support useful applications for small businesses.</p> <p>I believe developers can create meaningful technology by focusing on everyday problems within their communities.</p> <p><strong>Project:</strong> ListingWriter</p> <p><strong>Developer:</strong> Jungudo Muhammad Tukur</p> <p><strong>GitHub:</strong> <a href="https://github.com/hafeezmt" rel="noopener noreferrer">https://github.com/hafeezmt</a></p> mlhacks devchallenge hackathon Jungudo Muhammad Tukur Thu, 15 Oct 2026 13:50:22 +0000 https://dev.to/hafeezmt/listingwriter-i-built-a-private-ai-property-listing-generator-for-nigerian-real-estate-agents-3ffd https://dev.to/hafeezmt/listingwriter-i-built-a-private-ai-property-listing-generator-for-nigerian-real-estate-agents-3ffd <h1> ListingWriter: AI-Powered Property Marketing for Nigerian Real Estate Agents </h1> <h2> What I Built </h2> <p>I built <strong>ListingWriter</strong>, an open-source AI application designed to help Nigerian real estate agents create professional property marketing content quickly and easily.</p> <p>Instead of writing separate descriptions for different social media platforms, agents can enter their property details once and generate:</p> <ul> <li>A complete professional property listing</li> <li>A short WhatsApp broadcast message</li> <li>An Instagram caption with relevant hashtags</li> </ul> <p>ListingWriter supports <strong>English, Nigerian Pidgin, and Hausa (beta)</strong>, making it more relevant to Nigeria's diverse property market.</p> <h2> The Problem </h2> <p>Real estate agents frequently advertise properties across WhatsApp, Instagram, and other platforms. Preparing different descriptions for each channel takes time, and paid AI services can become expensive.</p> <p>ListingWriter aims to make property marketing faster, more affordable, and accessible to smaller real estate businesses.</p> <h2> How It Works </h2> <p>ListingWriter uses <strong>Python, Ollama, and Google's open-weight Gemma 3 model</strong>.</p> <p>The user provides property information, selects their preferred language, and generates marketing content using AI running locally on their computer.</p> <p>The application uses the lightweight <code>gemma3:1b</code> model by default.</p> <h2> Why Open-Source AI Matters </h2> <p>I chose a locally hosted AI model for three important reasons:</p> <p><strong>1. Lower operating costs</strong></p> <p>Agents can generate listings without paying per-request AI API fees. This creates opportunities for a more affordable service.</p> <p><strong>2. Privacy</strong></p> <p>Local AI processing means property and client information does not need to be sent to an external AI provider.</p> <p><strong>3. Offline functionality</strong></p> <p>After downloading and configuring the model, agents can generate content without a continuous internet connection.</p> <p>This is especially useful in locations where internet connectivity is unreliable.</p> <h2> The Side Hustle Opportunity </h2> <p>ListingWriter could become an affordable marketing tool for independent real estate agents and agencies.</p> <p>My proposed pricing model is:</p> <ul> <li> <strong>Free:</strong> Five listings per month</li> <li> <strong>Agent:</strong> ₦5,000/month for unlimited listings</li> <li> <strong>Agency:</strong> ₦15,000/month for up to five agents</li> </ul> <p>These are proposed subscription prices, not an existing paid service.</p> <h2> Challenges and Lessons </h2> <p>Local AI introduces trade-offs. Smaller models can produce inconsistent translations, especially in Nigerian Pidgin and Hausa. Generating content can also be slower on computers without dedicated GPUs.</p> <p>This project helped me explore how open-weight models can support practical business applications without depending on commercial AI APIs.</p> <h2> Demo </h2> <p>Watch my ListingWriter AI project demonstration:</p> <p> <iframe src="https://www.youtube.com/embed/UOiOf1euKgE" width="710" height="399"> </iframe> </p> <h2> Source Code </h2> <p><strong>GitHub:</strong> <a href="https://github.com/hafeezmt/listingwriter" rel="noopener noreferrer">https://github.com/hafeezmt/listingwriter</a></p> <p>The repository contains the Python application, dependencies, setup instructions, and MIT license.</p> <h2> What's Next? </h2> <p>I plan to improve ListingWriter by adding:</p> <ul> <li>Better Hausa and Yoruba support</li> <li>Listing history</li> <li>Image-Based social media exports</li> <li>More reliable and consistent generated descriptions</li> </ul> <h2> Final Thoughts </h2> <p>ListingWriter demonstrates how open-source AI can help Nigerian entrepreneurs solve everyday business problems.</p> <p>My goal is to make useful AI technology accessible to small businesses, particularly in communities where affordability, privacy, and connectivity matter.</p> <p><strong>Built by Jungudo Muhammad Tukur</strong></p> <p>GitHub: <a href="https://github.com/hafeezmt" rel="noopener noreferrer">https://github.com/hafeezmt</a></p> devchallenge hf26challenge opensource ai