cyberrscourse

Posted on Sep 5

CharityCheck AI - Verify Charities with AI-Powered Transparency Analysis

#weekendchallenge #webdev #productivity

DEV Weekend Challenge: Generosity Edition Submission 💜

Description: An AI-powered tool that helps donors make informed decisions by analyzing charity transparency, detecting red flags, and providing personalized recommendations using Google Gemini.

This is a submission for Weekend Challenge: Generosity Edition

What I Built

CharityCheck AI is an AI-powered charity transparency and verification tool that helps donors make informed giving decisions.

The problem: 82% of donors want to verify charities before giving, but financial reports are complex and time-consuming to analyze. Most people donate based on emotional appeals without understanding where their money actually goes.

CharityCheck AI solves this by:

Searching charities and displaying key financial metrics (program spending, overhead, ratings)
AI-powered analysis using Google Gemini to detect red flags, assess trust, and explain impact
Personalized recommendations matching donors with verified charities based on their interests

Think of it as "RateMyProfessor for charities" — instant transparency powered by AI.

Key Features

🔍 Charity Search & Verification

Financial health score
Accountability ratings
Program spending percentage (how much goes to actual cause vs overhead)
Fundraising efficiency metrics

🤖 AI Transparency Analysis (powered by Google Gemini)

Trust score (0-100) based on all metrics
Red flag detection (high overhead, low accountability, etc.)
Impact analysis: what your $100 donation achieves
Plain-English explanations of complex financial data

💡 Personalized Charity Matching

Describe causes you care about
AI recommends top 3 verified charities with explanations
Get next steps: how to donate, volunteer, or get involved

Demo

Live Demo

🚀 https://charitycheckai.onrender.com/

Screenshots

Search & Analysis

Search any charity and get instant AI-powered transparency analysis

AI Recommendations

Get personalized charity recommendations based on your interests

Video Demo

📹

How It Works

User searches "American Red Cross"
System displays financial metrics (89% program spending, 3/4 stars)
Google Gemini AI analyzes the data:
- Calculates trust score: 85/100
- Identifies strengths: "Strong accountability, transparent reporting"
- Detects concerns: "Above-average overhead (11%)"
- Shows impact: "$100 provides disaster relief supplies for 2 families"
User makes informed decision whether to donate

Code

Repository: https://github.com/cyberrscourse/charityAI
Note: Source code available for review. All development completed within the challenge window (Sept 4-7, 2026).

Tech Stack

Backend: Python Flask
AI Engine: Google Gemini API
Frontend: HTML, CSS, JavaScript (vanilla)
Data: Charity ratings based on public financial data methodology

Quick Start

pip install flask flask-cors google-genai
export GOOGLE_API_KEY="your_key_here"
python app.py

How I Built It

Problem Definition

82% of donors want to verify charities before giving, but charity financial reports are complex (100+ page IRS 990 forms) and verification sites are often paywalled. The gap: people need simple, trustworthy charity transparency insights.

Technical Architecture

User Input → Flask API → Google Gemini → Analyzed Insights
                ↓
          Charity Database

Design decisions:

Flask: Lightweight, fast deployment
Google Gemini: Natural language understanding for financial analysis
No database: In-memory storage for demo (production would use Charity Navigator API)
Vanilla JS: Zero build complexity, deploy anywhere

Google AI Integration (Prize Category Focus)

Google Gemini powers three core features:

1. Financial Transparency Analysis

Gemini analyzes charity financial metrics and generates human-readable assessments:

Input:

Financial health score
Program spending percentage
Fundraising efficiency
Accountability rating

AI Processing:

prompt = f"""
Analyze this charity's financial data and provide:
1. Trust Score (0-100)
2. Key Strengths (2-3 bullet points)
3. Red Flags (if any exist)
4. Donation Recommendation
5. Impact Analysis: What $100 achieves

Charity: {charity['name']}
Financial Score: {charity['financial_score']}/100
Program Spending: {charity['program_percentage']}%
Fundraising Cost: ${charity['fundraising_efficiency']} per $1 raised
"""

Output Example:

Trust Score: 96/100

Key Strengths:
• Perfect accountability score (100/100)
• Exceptional program spending (100% to programs)
• Zero fundraising costs

Recommendation: Highly recommended - best-in-class transparency

Impact: $100 = 100% to clean water projects in developing countries

Why AI is essential: Gemini understands financial context (85%+ program spending = excellent, <70% = red flag) and explains why numbers matter, not just what they are.

2. Personalized Charity Matching

Semantic Understanding:

User: "I care about clean water and helping children"

AI matches based on intent, not keywords:

Charity Water (water access) - 4/4 stars, 100% program spending
Doctors Without Borders (children's healthcare) - 4/4 stars, 88% program spending
UNICEF (child welfare) - aligned with interests

Traditional keyword matching would miss nuanced connections. Gemini understands "helping children" includes healthcare, education, and nutrition charities.

3. Red Flag Detection

Gemini identifies concerning patterns across multiple metrics:

High overhead (>25%) + low rating → "Inefficient fund use"
Low accountability + poor transparency → "Limited financial disclosure"
Recent negative coverage → "Controversies detected"

This would require 100+ lines of rule-based code. Gemini does it in one API call.

Technical Wins

✅ Structured prompting: Consistent AI output format

✅ Progressive disclosure: Show data first, AI analysis second (better UX)

✅ Zero authentication: No login friction

✅ Error handling: Graceful failures with actionable messages

Without AI vs. With AI

❌ Without AI: "Program Percentage: 87.9%"

✅ With AI: "88% of donations fund medical care in crisis zones — your $100 treats 12 patients"

The difference: AI makes financial data accessible to non-experts.

Prize Categories

Best Use of Google AI

This project demonstrates Google Gemini's strengths:

Natural Language Understanding - Semantic charity matching based on user intent
Financial Data Analysis - Contextual interpretation of complex metrics
Explainability - Doesn't just recommend, explains why
Pattern Recognition - Multi-metric red flag detection
Accessibility - Translates technical financial data into plain English

Technical Implementation:

Structured prompts for consistent output
Multi-step AI workflows (search → analyze → recommend)
Context-aware responses (charity type matters for benchmarks)

Impact

Use Cases

Donors: Verify charities before giving
Nonprofits: Understand how transparency affects public perception
Educators: Teaching financial literacy through real charity data
Researchers: Analyzing charity sector transparency trends

Next Steps

Phase 1 (Weeks 1-2):

Charity Navigator API integration (200k+ charities)
Data export functionality

Phase 2 (Month 1-3):

Historical financial trend analysis
International charity support
Public API for third-party integrations

Focus: Data accuracy and actionable insights.

Why This Matters

Charitable giving is a $500B+ global industry with limited consumer protection. Bad actors waste billions; good charities struggle to prove trustworthiness.

CharityCheck AI democratizes transparency.

By combining public financial data with AI analysis:

✅ Donors make informed decisions
✅ Transparent charities get rewarded with more donations
✅ Giving becomes more effective

Better transparency → more trust → more generosity.

Built for: DEV Weekend Challenge: Generosity Edition

Stack: Google Gemini AI, Python Flask, HTML/CSS/JS

Time: 48 hours

Code: ~500 lines

Hosted on Render free tier - first load may take ~50 seconds as the instance spins up. Subsequent requests are instant.

DEV Community