Finformo

loading
Credibility Indicator
Reason model
Reason Model

Inspiration

We initially had the idea of creating an extension that would help with researching, functioning similarly to what we have now in that an overview would be generated with details on the viability of the information and creators of an article. Finding viable information on the internet is a daunting and unclear task that requires its own dedicated time and effort. Helping individuals be more keen on the context of the information they consume by bringing details like this to them directly and automatically is a good step in preventing misinformation. The challenge presented by OneEthos, however, steered us into molding this idea with financial advising in mind as the forefront. After all, one of the most important things that people ought to be well-informed about is their money. This sort of analyzer helps people think twice before choosing to believe a possibly sleazy finance article that provides nothing but bad advice.

What it does

Finformo activates when the user opens an online financial article and clicks the extension icon. It then runs a comprehensive analysis of the article's credibility by evaluating multiple factors: the publisher's reputation (using a curated database of 40+ financial news sources), content quality indicators (emotional language, source balance, factual accuracy), and transparency issues. The extension displays an interactive popup with a credibility score (0-100), fact-check ratings, article status, and detailed bias indicators. Users can explore tabs for in-depth analysis, publisher information, and source verification—all presented in an intuitive interface.

How we built it

Front-End:

Constructed a manifest.json file outlining the metadata of our Chrome extension following Manifest V3 specifications
Created popup.html with a multi-tab interface displaying credibility overview, analysis indicators, and publisher information
Designed styles.css with a custom color palette (green, purple, grey, white) featuring smooth animations and responsive UI elements
Built popup.js to handle article content extraction, API communication, and dynamic result rendering
Implemented content.js for webpage integration

Back-End:

Developed a serverless API using Vercel with Node.js
Integrated Google's Gemini 2.5 Flash API for AI-powered content analysis
Created a comprehensive source credibility database (sources.js) ranking 40+ financial publishers across 5 tiers
Built classify.js endpoint handling CORS, request validation, and multi-factor credibility scoring
Implemented weighted scoring algorithm combining AI analysis (60%) with source reputation (40%)
Deployed backend infrastructure with environment variable management for secure API key storage

Challenges we ran into

Network Configuration: Navigated CORS header setup and environment variable management across local development and production environments
Template Literal Syntax: Debugged JavaScript syntax errors with nested template literals and proper quote usage in prompts
Extension Architecture: Understanding Chrome extension structure, manifest requirements, and the interaction between popup scripts and content scripts

Accomplishments that we're proud of

Successfully built a fully functional Chrome extension from concept to deployment in a hackathon timeframe
Achieved seamless UI integration with a polished, professional interface featuring smooth animations and intuitive navigation
Successfully integrated Gemini 2.5 Flash API for sophisticated financial content analysis
Created a comprehensive source credibility system with tiered rankings spanning government sources to questionable outlets
Implemented a robust multi-factor scoring algorithm that balances AI analysis with source reputation
Overcame significant technical hurdles with WSL, Vercel deployment, and API configuration
Delivered an extension that provides genuine value in combating financial misinformation

What we learned

How to design and build Chrome extensions from scratch, including manifest configuration, popup architecture, and content script integration
Integrating and prompting Google's Gemini API for structured content analysis and JSON response formatting
Deploying serverless functions on Vercel with proper environment variable management and CORS handling
Working with ESM modules in Node.js and managing package dependencies
Balancing AI-powered analysis with rule-based credibility scoring for more reliable results
The importance of source verification and multi-dimensional analysis in fact-checking applications

What's next for Finformo

Expand Beyond Finance: Transform Finformo into a general-purpose fact-checker that can analyze articles across all topics—politics, health, science, technology—making credible information accessible to everyone
Enhanced Source Analysis: Implement web scraping to automatically verify claims against authoritative sources and fact-checking databases
Real-Time Fact Verification: Integrate with fact-checking APIs (Google Fact Check, PolitiFact, Snopes) to cross-reference specific claims
Citation Tracking: Analyze and verify the quality of sources cited within articles
User Feedback System: Allow users to report inaccuracies and contribute to improving the credibility database
Browser Compatibility: Extend support to Firefox, Safari, and Edge
Machine Learning Improvements: Train custom models on verified misinformation datasets for more accurate detection
Community Features: Build a platform where users can share credibility reports and discuss questionable content