veritas-engine

• 

  home page

• 

  execution first half

• 

  execution second half

Inspiration

The internet moves faster than truth. News spreads instantly, but verification still takes time and expertise. We built Veritas Engine to close that gap—giving users a fast, AI-driven way to assess the credibility of news articles before misinformation takes hold.

What it does

Frontend: Next.js (App Router) with Tailwind CSS for a clean, responsive UI

Backend logic: Custom analysis pipeline using Gemini AI for claim extraction and reasoning

Architecture: URL-based execution flow with real-time terminal-style feedback

UX focus: Transparent analysis logs + final verdict instead of a black-box score

The system is designed to feel like an “engine” rather than a simple classifier.

How we built it

Frontend: Next.js (App Router) with Tailwind CSS for a clean, responsive UI

Backend logic: Custom analysis pipeline using Gemini AI for claim extraction and reasoning

Architecture: URL-based execution flow with real-time terminal-style feedback

UX focus: Transparent analysis logs + final verdict instead of a black-box score

The system is designed to feel like an “engine” rather than a simple classifier.

Challenges we ran into

Handling routing and execution flow correctly in Next.js (App Router quirks)

Designing analysis output that is explainable, not just a verdict

Preventing UI dead-ends (404s, broken state transitions)

Balancing speed with meaningful AI reasoning

Every issue forced us to tighten both logic and architecture.

Accomplishments that we're proud of

A working end-to-end AI credibility analysis pipeline

Real-time analysis feedback instead of static loading screens

A professional, product-grade UI suitable for demos and scaling

Clean separation between execution, analysis, and presentation

This isn’t a demo toy—it’s a foundation.

We learned to think like product engineers, not just coders.

What we learned

AI systems must explain why, not just answer what

Frontend architecture matters as much as model quality

Small routing mistakes can break entire experiences

Building trust-focused products requires clarity over cleverness

We learned to think like product engineers, not just coders.

What's next for veritas-engine

Multi-source cross-verification and citations

Credibility scoring with confidence ranges

Browser extension for instant article checks

User feedback loops to improve verdict accuracy

Dataset-backed evaluation and benchmarking

Goal: Make Veritas Engine a daily-use tool for fighting misinformation.

Built With

• esmodules
• gemini3pro
• googleaistudios
• googlefonts
• googlegemini
• googlegeminiapi
• lucidereact
• react19
• tailwindcss