WAE Viral Prediction Engine v1.1
Model Description
WAE Viral Prediction Engine is an AI-native media intelligence model that predicts content virality scores (0-100) before publishing. It is the first module of the Brand Factory Intelligence OS — a comprehensive system designed to transform how brands create, optimize, and distribute content.
Built on 833M+ first-party organic views from 2,720 real posts primarily documented through official Meta/Facebook/Instagram analytics exports in the wildlife/nature content vertical.
Current public evidence archive primarily documents Meta/Facebook/Instagram records. Additional platforms may be added in future versions.
Note: This is a hosted API model. No weights or source code are distributed. Access is via the WAE API.
Key Capabilities
| Capability | Description |
|---|---|
| Viral Score Prediction | Score any content caption 0-100 with tier classification |
| Content Generation | Generate optimized video prompts, captions, and hashtags |
| Content Analysis | Deep-dive analysis of why content performs or underperforms |
| Viral Trigger Detection | Identify psychological triggers that drive sharing |
Model Architecture
Input Layer (21 features)
├── Description Quality Score (NLP)
├── Viral Trigger Count & Type
├── Temporal Optimization Score
├── Platform-Specific Signals
├── Content DNA Markers
└── Engagement Probability Vectors
│
┌────┴────┐
│ Ensemble │
│ Layer │
├──────────┤
│ Random │ Gradient │
│ Forest │ Boosting │
│ (500 │ (500 │
│ trees) │ rounds) │
└────┬─────┴─────┬─────┘
│ │
┌────┴───────────┴────┐
│ Weighted Average │
│ (RF: 0.6, GB: 0.4) │
└──────────┬───────────┘
│
Output: Viral Score (0-100) + Tier
Training Data
| Metric | Value |
|---|---|
| Total Posts Analyzed | 2,720 |
| Total Organic Views | 833,000,000+ |
| Platforms Covered | Facebook, Instagram (Meta official exports) |
| Content Vertical | Wildlife / Nature / Animals |
| Time Period | 2020-2025 |
| Features Extracted | 21 per post |
Performance Metrics
| Metric | Score |
|---|---|
| Tier Classification Accuracy | 86.2% |
| Mean Absolute Error (MAE) | 8.3 points |
| R² Score | 0.74 |
| Top-Tier Precision | 89.1% |
| Top-Tier Recall | 82.4% |
Usage (API)
curl -X POST https://ai.wildanimalencounter.com/api/score \
-H "Content-Type: application/json" \
-d '{
"animal": "Lion",
"description": "A majestic lion roaring at sunset in the Serengeti",
"platform": "instagram"
}'
Limitations
- Currently trained on wildlife/nature content only
- Text/metadata analysis only (no video frame analysis yet)
- English-optimized
- Platform algorithm changes may affect predictions over time
Links
- Live Demo: https://ai.wildanimalencounter.com
- API Documentation: https://ai.wildanimalencounter.com/api-docs
- Benchmark Report: https://ai.wildanimalencounter.com/benchmark
- Company: https://wildanimalencounter.com
Built by WAE Media — AI-Native Media Intelligence