Biblical Prophecy Intelligence Platform - Backend

Complete Cloudflare Worker backend with Hugging Face AI scoring for prophecy news analysis.

Features

🤖 AI-Powered Scoring

Hugging Face Integration: Uses facebook/bart-large-mnli for zero-shot classification (FREE)
Multi-factor Analysis: Combines AI, keyword matching, and target individual detection
Smart Threshold: Only saves articles with score ≥ 5.0

📡 RSS Feed Processing

Hybrid Approach: RSS parsing with web scraping fallback
Auto-scheduling: Runs every 15 minutes via Cron Triggers
OPML Import: Bulk import 4000+ feeds at once
Error Handling: Tracks failures and retries

🎯 Target Monitoring

15 Pre-loaded Individuals: Macron, Schwab, Netanyahu, etc.
Tier System: Tier 1 (high priority) vs Tier 2 (medium)
Mention Tracking: Automatically detects and weights mentions

📊 Sophisticated Scoring Algorithm

Overall Prophecy Score = (AI × 0.4) + (Keywords × 0.35) + (Targets × 0.25)

1. AI Score (40% weight)

Classifies text against 12 prophecy-related categories
Falls back to keyword-based analysis if API unavailable

2. Keyword Score (35% weight)

Weighted keywords:

Critical (2.0): antichrist, mark of the beast, third temple
High (1.5): tribulation, rapture, armageddon, Ezekiel 38
Medium (1.0): prophecy, Israel, Jerusalem, digital currency
Low (0.5): biblical, Christian, church, faith

3. Target Score (25% weight)

Tier 1 individuals: 3.0 points per mention
Tier 2 individuals: 2.0 points per mention
Additional keyword matches: 0.3 points

📁 File Upload Support

Documents: PDF, DOC, DOCX, TXT
YouTube: URL processing (transcript extraction ready)
Manual Entry: Direct article submission

API Endpoints

Health & Stats

GET  /api/health              → Health check
GET  /api/stats               → System statistics

Feed Management

GET    /api/feeds             → List all feeds
POST   /api/feeds             → Add new feed
PUT    /api/feeds/:id         → Update feed
DELETE /api/feeds/:id         → Delete feed
POST   /api/feeds/opml        → Import OPML file

Articles

GET  /api/articles/recent?limit=50&minScore=5.0  → Recent articles
POST /api/articles/score      → Score single article

Ingestion

POST /api/ingest              → Trigger RSS scan
POST /api/ingest/single       → Submit single article

Targets

GET  /api/targets             → List all monitored individuals
POST /api/targets             → Add new target

Uploads

POST /api/upload/document     → Upload PDF/DOC/TXT
POST /api/upload/youtube      → Submit YouTube URL

Dashboards

GET  /api/dashboard/antichrist   → Antichrist site data
GET  /api/dashboard/prophecy     → General prophecy site data

Setup Instructions

1. Install Wrangler CLI

npm install -g wrangler

wrangler login

3. Create D1 Database

wrangler d1 create prophecy-intelligence

Copy the database_id from the output and update wrangler.toml:

database_id = "xxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx"

4. Initialize Database Schema

wrangler d1 execute prophecy-intelligence --file=../YellowKidokc-enhanced-prophecy-dashboard-bigcards-frontend/schema.sql

This will:

Create all tables (analyzed_articles, rss_sources, target_individuals, etc.)
Add indexes for performance
Pre-load 15 monitored individuals (Tier 1 & 2)
Pre-load 68 RSS feeds from your OPML

5. Get Hugging Face API Key (FREE)

Go to https://huggingface.co/settings/tokens
Create account (free)
Generate new token with “Read” access
Set as secret:

wrangler secret put HUGGINGFACE_API_KEY

When prompted, paste your token.

6. Deploy to Cloudflare

wrangler deploy

You’ll get a URL like: https://prophecy-intelligence-backend.your-subdomain.workers.dev

7. Test the API

curl https://your-worker-url.workers.dev/api/health

Expected response:

{
  "ok": true,
  "status": "healthy",
  "timestamp": "2025-..."
}

8. Trigger Initial Ingestion

curl -X POST https://your-worker-url.workers.dev/api/ingest

This will start processing all active RSS feeds.

Importing 4000 RSS Feeds via OPML

If you have an OPML file with 4000 feeds:

curl -X POST https://your-worker-url.workers.dev/api/feeds/opml \
  -H "Content-Type: application/xml" \
  --data-binary @your-feeds.opml

Response:

{
  "ok": true,
  "imported": 4000
}

Environment Variables

Set via wrangler secret:

HUGGINGFACE_API_KEY - Your HuggingFace API token (required for AI scoring)

Scheduled Ingestion

The worker automatically runs every 15 minutes to:

Fetch all active RSS feeds
Parse new articles
Score each article
Save articles with score ≥ 5.0
Update feed statistics
Track errors

View logs:

wrangler tail

Testing Locally

wrangler dev

This starts a local server at http://localhost:8787

Test scoring:

curl -X POST http://localhost:8787/api/articles/score \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Emmanuel Macron Proposes Seven-Year Peace Covenant for Middle East",
    "content": "French President Emmanuel Macron announced today a comprehensive seven-year peace plan involving Israel, Palestine, and regional partners. The covenant includes provisions for joint governance of Jerusalem and reconstruction of a worship site on the Temple Mount."
  }'

Expected high score (8-10) due to:

Target mention: Emmanuel Macron (Tier 1 → 3.0 points)
Keywords: “seven-year” (2.0), “peace covenant” (2.0), “Temple Mount” (2.0), “Jerusalem” (1.0)
AI classification: High prophecy relevance

Database Schema

Tables:

analyzed_articles - Scored articles (only ≥ 5.0)
rss_sources - Feed configurations
target_individuals - Monitored people (15 pre-loaded)
article_target_mentions - Article-to-target relationships
keyword_matches - Keyword tracking per article
system_alerts - High-priority notifications
analysis_sessions - Ingestion run logs
export_logs - Data export history

Architecture

┌─────────────────────────────────────────────┐
│  Cloudflare Worker (Serverless)             │
│  ├── RSS Feed Fetcher                       │
│  ├── Web Scraper (fallback)                 │
│  ├── Hugging Face AI Client                 │
│  ├── Scoring Engine (multi-factor)          │
│  ├── D1 Database Interface                  │
│  └── CORS-enabled REST API                  │
└─────────────────────────────────────────────┘
           ↓                    ↓
  ┌─────────────────┐   ┌──────────────────┐
  │  Cloudflare D1  │   │  Hugging Face    │
  │  (SQLite)       │   │  Inference API   │
  │  - Articles     │   │  (FREE tier)     │
  │  - Feeds        │   └──────────────────┘
  │  - Targets      │
  └─────────────────┘

Cost

FREE (within Cloudflare free tier limits):

100,000 Worker requests/day
5 million D1 reads/day
100,000 D1 writes/day
Hugging Face Inference API free tier

For 4000 feeds @ 15-min intervals:

4000 feeds × 4 checks/hour × 24 hours = 384,000 requests/day
You’ll need Cloudflare Workers Paid plan ($5/month) for this volume
Or reduce check frequency to hourly (96,000 requests/day → FREE)

Troubleshooting

”Database not found"

wrangler d1 list
wrangler d1 execute YOUR_DB_NAME --file=schema.sql

"Hugging Face API returns 503”

Model is loading (cold start). Retry in 30 seconds or use fallback scoring (automatic).

”No articles being saved”

Check scores are ≥ 5.0:

curl -X POST https://your-worker.workers.dev/api/articles/score \
  -d '{"title":"Test","content":"antichrist temple mount israel prophecy"}'

“CORS errors”

CORS headers are included in all responses. Check browser console for actual error.

Next Steps

Deploy Frontend: Use admin.html for feed management
Create Two Sites:
- Antichrist site: /api/dashboard/antichrist
- Prophecy site: /api/dashboard/prophecy
Set up Monitoring: Use Cloudflare Analytics dashboard
Optimize Scores: Adjust weights in scoreArticle() function
Add More Sources: Import your 4000-feed OPML

Support

Issues? Check:

Cloudflare dashboard for errors
wrangler tail for real-time logs
D1 console for database queries

License

MIT

Canonical Hub: CANONICAL_INDEX

GO Vault

Explorer

README