Why I Created Obtainabus
Why I Created Obtainabus
Buying the right product should be straightforward. In reality, it often means opening countless tabs, reading specifications that are difficult to compare, sorting genuine differences from marketing language, and wondering whether a deal is actually good value.
I created Obtainabus to make that process simpler.
Obtainabus is an AI-powered shopping assistant. You tell it what you are looking for in ordinary language, including your budget and the things that matter to you, and it researches suitable products from retailers in your market. It then explains the strongest options, their advantages and disadvantages, and the trade-offs between them.
You might ask for a quiet robot vacuum that can cope with pet hair, a dependable laptop for work, or a television that performs well in a bright room. Obtainabus is designed to understand the need behind the purchase, rather than treating a search as a collection of disconnected keywords.
From Laptop Copilot to Obtainabus
The project began with a much narrower idea called Laptop Copilot. Laptops are particularly difficult to compare because model names, processors, memory, screen quality and battery claims can quickly become overwhelming. I wanted to build something that could translate all of that into useful advice for a real person.
Once the first version was working, it became clear that the same problem existed far beyond laptops. People face similar confusion when choosing phones, televisions, headphones, monitors, appliances, fitness equipment and many other products. I expanded the search system to understand different product categories and renamed the project Obtainabus.
That change was more than a new name. A useful shopping assistant cannot compare every type of product in exactly the same way. Battery life may matter greatly for a laptop, while suction, noise and floor type matter for a vacuum cleaner. I built category-aware searches and specifications so that the app focuses on the details that are relevant to each purchase.
What Obtainabus does
The experience starts as a conversation. A person describes what they need and Obtainabus can ask follow-up questions, recognise conflicting requirements and refine the search. If someone asks for an “ultra-portable 17-inch laptop,” for example, the assistant can point out the tension between those two priorities rather than silently returning unsuitable results.
It searches live shopping information, groups comparable products and filters out irrelevant accessories. The results are then ranked against the person’s stated needs. Instead of presenting unexplained scores, Obtainabus gives plain-English reasons, useful specifications, pros and cons, review patterns and a summary of the main compromises.
People can compare products side by side, save useful results and return to earlier research. Obtainabus can also retain preferences and learn from feedback and purchases, helping later recommendations reflect what the person has found useful before. Market and currency support means the search is scoped to where somebody actually shops, rather than merely converting a price from another country.
The aim is not to make the final decision for anyone. It is to do the repetitive research, organise the evidence and make the decision easier to understand. The person remains in control of which trade-offs matter most.
How I built it
I developed Obtainabus as one connected system with a web application, mobile apps and a shared service behind them. The web experience uses React, while the iPhone and Android apps are built with React Native and Expo. The server uses TypeScript, Express and PostgreSQL.
Several services work together behind the scenes. Live shopping search supplies current products and retailer information. Product-data services help fill in technical specifications. AI interprets the request, asks useful questions, ranks the results and explains why each recommendation may or may not fit. Authentication keeps saved research tied to the right account, and the subscription system supports Obtainabus Pro.
One of the most important technical changes was splitting search into two stages. Early versions waited for product search and full AI analysis to finish before showing anything, which could leave a person staring at a loading screen for too long. I redesigned the process so that relevant products appear first and the deeper ranking follows. That made the app feel much faster without giving up the quality of the analysis.
I also added safeguards for the times when an outside service is slow or unavailable. Searches use time limits and retries, and the ranking system has a dependable fallback. Product grouping prevents unrelated categories from becoming mixed together, while category-specific filters stop accessories from appearing as if they were the main product. These details are not especially visible, but they make the difference between an interesting demonstration and something people can rely on.
Building for the web and mobile
After establishing the web version, I wanted Obtainabus to be useful at the moment people actually shop. I built native-style apps for iPhone and Android, sharing the same accounts, searches and saved information with the website.
That brought a different set of challenges. Sign-in had to work consistently across browsers and two mobile platforms. Subscriptions needed to respect the different requirements of the web, Apple and Google. A conversation had to survive navigation and interruptions without losing its context. I worked through those problems one by one, testing the complete journey rather than treating each screen as an isolated feature.
Obtainabus is now available through its website, the Apple App Store and Google Play.
What I learned while creating it
The project taught me that AI is most useful when it is part of a carefully designed process. A clever answer is not enough. The underlying products must be current, comparable and relevant. The advice needs to show its reasoning. The system has to handle slow services, incomplete specifications, changing prices and the many ways in which people describe what they want.
I also learned the value of building in stages. Obtainabus started with one product category and a clear problem. Each later feature—conversational refinement, comparisons, price information, saved research, mobile apps and personalisation—was added to solve something I had observed in the experience. That gradual approach allowed the original idea to grow without losing sight of its purpose.
I am still developing Obtainabus. Retail information changes, AI improves and every type of purchase raises new questions. My goal remains the same as it was at the beginning: to reduce the time and uncertainty involved in product research, explain the things that genuinely matter, and help people buy with greater confidence.
