Project Overview
With digital content multiplying by the second and misinformation spreading faster than facts, staying accurately informed has become a real challenge. A forward-thinking media-tech startup approached Cubet with an ambitious vision: develop a mobile-first news platform that doesn't just deliver news, but evaluates its credibility in real time. The mission was to empower users to make informed choices by filtering out misleading or low-quality stories, using AI as the backbone of trust. The goal wasn’t to build just another news app. It was to build a platform where information is verified, categorised, and served with transparency, so readers always know what they’re consuming and why it matters.
Industry
Media-Tech/Mobile News Aggregation/AI in Journalism
The Client
A rising media-tech startup determined to restore trust in digital news consumption. Focused on transparency, credibility, and user empowerment, the client was looking to stand apart from traditional aggregators by integrating advanced fake news detection from the ground up.
Challenges Addressed
The client came to Cubet with several clear, but technically demanding, objectives.
- Information Overload: Users are bombarded with news, but lack clarity on what's real or reliable.
- Fake News Epidemic: Misinformation travels fast, often faster than fact-checked journalism.
- Credibility Gap: Even trusted outlets can have bias; users need a tool that helps them assess source trust.
- Cross-Platform Consistency: Deliver a high-performance, intuitive app on both Android and iOS.
- Real-Time Delivery: Ensure the backend can handle live news aggregation and instant scoring.
Each challenge called for a thoughtful, tech-enabled solution that could scale with demand while preserving performance, speed, and reliability.
Collaboration in Action
Cubet partnered with the startup to bring the concept to life, right from design and architecture to development and deployment. The result? A robust, AI-powered news app that not only aggregates articles from hundreds of global sources but intelligently scores each article based on trust indicators such as tone, bias, source credibility, and sensationalism.
How It Works:
- News is pulled using global RSS feeds, APIs, and media scrapers.
- Articles are routed through a custom-built NLP engine that assesses bias, narrative consistency, tone, and exaggeration heuristics.
- Each story is then scored and labeled—"Verified", "Caution", or "Unverified"—based on a credibility scale.
- Users can tailor their feeds based on topics, regions, and trust filters, giving them greater control over their news experience.
The app doesn't just surface what’s trending, it tells you how reliable it is, and why.
Technologies Deployed
Mobile Frontend:
- React Native for seamless cross-platform compatibility
- Redux for efficient state management
- Push Notification System for verified or trending alerts
Backend Infrastructure:
- Node.js with Express to deliver lightweight RESTful APIs
- MongoDB to manage articles, user profiles, and metadata
- Redis to power low-latency real-time feed delivery
AI & NLP Engine:
- Built on spaCy and BERT, the system parses content for bias, polarity, tone, and keyword density
- Source reputation scoring using domain-level metadata
- Continuous learning loop incorporating flagged content and user feedback to fine-tune the detection model
Cloud & DevOps:
- AWS-hosted with integrated CDN for global performance
- Firebase for device sync and secure user authentication
- Cron-based aggregators fetching hourly updates from multiple sources
Innovative Feature
The standout innovation is the real-time fake news detection engine. It goes beyond surface-level filters and evaluates deeper content features: emotional charge, sentence structure, and headline patterns. Combined with domain reputation analysis and continuous learning from user inputs, the app becomes smarter every day, delivering more accurate assessments the more it's used. This is what transforms a simple news reader into an intelligent, trustworthy media assistant.
Value Delivered
The collaboration between Cubet and the client led to several measurable and strategic wins:
- Cross-platform launch on both Google Play and Apple App Store
- Over 100,000 articles processed monthly, all scored and tagged in real time
- Boost in average session times, indicating greater user engagement through credibility-tagged stories
- Adoption by early users in education and journalism, who value news transparency
- Improved content quality, as the algorithm de-emphasises clickbait and surfaces high-integrity stories
The platform has become more than a news app. It’s a daily tool for critical thinking.
User Feedback
Initial feedback has been overwhelmingly positive. Users particularly appreciated:
- The credibility tagging, which makes it easier to trust or question a story
- The ability to customise news feeds based on interest and trust levels
- A clean, fast, intuitive interface that works equally well on Android and iOS
Journalism educators, media professionals, and tech-savvy news consumers alike have highlighted the app's value in filtering noise and surfacing real, verified content.
Conclusion
Cubet turned a bold idea into a reality, proving that trust and technology can work hand in hand. This wasn’t just an app build; it was a tech intervention in how society consumes news. With a unique blend of NLP, real-time aggregation, and thoughtful UX, the solution doesn't just inform, it empowers.
Impact Made
- Elevated user expectations for transparency in digital news
- Reduced exposure to misleading or fake content
- Encouraged critical media consumption habits
- Delivered a trusted platform in an untrustworthy information ecosystem
This is how technology should serve users, not by overwhelming them with data, but by helping them discern what truly matters.