Inspiration
Online shopping has become easier, but making the right purchasing decision is still difficult. Users often have to visit multiple websites, compare prices manually, check product specifications, investigate sellers, and understand complicated discounts before deciding what to buy.
A product with a lower displayed price is not always the best deal, and two listings that look identical may have completely different specifications. We wanted to solve this problem by building something beyond a traditional shopping app or product comparison tool.
That's how Valtix was born an AI-powered shopping intelligence platform designed to help users understand the complete shopping journey and make more confident purchasing decisions.
What We Built
Valtix brings product browsing, discovery, deal analysis, investigation, and decision-making together through two connected experiences: Valtix Web and Valtix Desktop Browser.
Its major features include:
- Smart Finder: Helps users discover relevant products based on their requirements, budget, and intended use.
- True Deal Engine: Breaks down displayed prices, applicable discounts, bank offers, and additional charges to help users understand the estimated payable amount.
- Variant Intelligence: Identifies whether similar-looking product listings actually match in specifications such as RAM, storage, and model generation.
- Purchase Investigator: Brings together available product, seller, price, and supporting evidence while highlighting missing information.
- Decision Cockpit: Provides a central workspace to compare products, analyse differences, and evaluate purchasing options.
- AI Research Assistant: Helps users understand product information, ask questions, explore trade-offs, and make informed decisions.
- Valtix Desktop Browser: Provides a dedicated browsing experience with multiple tabs and user-controlled product evidence capture.
- PayPal Sandbox Integration: Demonstrates a test checkout flow without processing real payments.
How We Built It
We built Valtix using modern web and desktop technologies.
The web application uses Next.js, React, and TypeScript, while the desktop browser is built using Electron with Chromium-based browsing capabilities.
We use Zod for structured data validation and Gemini AI for the research assistant.
One important part of our architecture is that we don't rely on AI for everything. Critical price calculations are handled by a deterministic pricing engine to keep financial calculations consistent and predictable.
We also use structured product information and evidence-based comparison logic to avoid treating different product variants as identical.
For payments, we integrated PayPal Sandbox to demonstrate the checkout experience in a test environment.
Challenges We Faced
One of our biggest challenges was identifying whether two product listings actually represented the same product. Similar names do not always mean identical specifications. We worked on product identity and variant-matching logic to avoid incorrect comparisons.
Another challenge was understanding the actual deal price. Discounts, bank offers, additional charges, and conditional offers can make price calculations complicated. We developed a structured pricing engine to handle these components separately.
Building the desktop browser and connecting it with the shopping intelligence experience also required careful handling of browser tabs, permissions, and evidence capture.
We also focused on keeping AI-generated research separate from critical financial calculations so that important decisions are not based on unsupported AI-generated numbers.
What We Learned
Building Valtix taught us that creating a useful AI-powered product is not just about integrating an AI model.
It requires good data structures, reliable calculations, thoughtful user experience, evidence handling, and clear boundaries around what the system can and cannot verify.
We learned how to combine AI with deterministic logic, build a desktop application alongside a web platform, and design features around real user problems.
What's Next?
We want to continue improving Valtix by expanding product intelligence, strengthening retailer integrations, improving evidence-based investigations, and making the browsing and shopping research experience even more connected.
Our vision is simple:
Find the right product. Understand the real deal. Make a confident decision.
Built With
- ai
- chromium
- electron
- fintech
- gemini
- javascript
- machine-learning
- next.js
- node.js
- paypal
- react
- shopping-intelligence
- tailwind
- typescript
- zod
