HackED 2025 Devpost Submission - "Bananalyze"
Introduction: The Problem We’re Solving
“LLMs are powerful, but they hallucinate, making them unreliable for research. What if AI could fact-check itself?”
AI-powered tools like ChatGPT have revolutionized education, research, and professional workflows, but misinformation remains a critical issue. LLMs generate the most probable response—not necessarily the most factual one.
The Problem Statement
AI-generated misinformation spreads unchecked, affecting students, journalists, researchers, and professionals who rely on it for credible information.
Our Solution: The Bananalyze Chrome Extension
We developed Bananalyze—a chrome extension that instantly extracts verifiable claims from any text, fact-checks them using trusted sources, and labels them as True or False with a sourced explanation.
Key Features
- Instant Fact-Checking – Just highlight text and click the popup to verify claims.
- Real-Time AI Analysis – Uses OpenAI’s LLM to extract and categorize claims.
- Trusted Sources – Pulls data from WHO, NIH, major news outlets, and scientific journals.
- Seamless UI – Works across any webpage with a simple, intuitive design.
- Transparency & Credibility – Shows source links, summaries, and trust ratings.
How It Works (Super Simple!)
- Highlight any text you want to fact-check.
- Click the pop-up banana indicator, and the extension does the rest.
- Get real-time results, including credibility ratings, explanations, and direct supporting sources.
Technical Breakdown
- AI Claim Extraction – Uses OpenAI to identify verifiable claims from user-highlighted text.
- Web Scraping & Source Verification – Pulls evidence from WHO, NIH, Google Scholar, and news sites.
- Claim Matching & Labeling – Program determines truthfulness based on verified data.
- Chrome Extension UI – Fetches instant results and displays them in a clean, user-friendly interface.
Tech Stack
- Frontend – JavaScript, HTML, CSS (Chrome Extension)
- Backend – Node.js, Express.js, Cors, dotenv, OpenAI
- AI/ML – OpenAI GPT (LLM-based claim extraction)
- Web Scraping – Google Custom Search API, Cheerio
- Storage – Chrome local storage
Scalability & Future Plans
- Expand Trusted Sources – Add peer-reviewed journals and fact-checking databases.
- Bias Detection – Identify partisan or misleading narratives in AI-generated responses.
- Academic Citation Checker – Verify citations in research papers and essays.
- API Integration – Allow other LLMs to fact-check their own responses before outputting them.
Built With
- cheerio
- chrome
- cors
- css
- dotenv
- express.js
- gpt
- html
- javascript
- node.js
- openai

Log in or sign up for Devpost to join the conversation.