| name | neural-hub |
| description | Gemini AI integration for market analysis, sentiment tracking, pattern recognition. 6 AI modules. (project) |
NEURAL HUB - GIO (Gemini) Integration
Location: /Volumes/LegacySafe/SS_III/neural_hub/
AI Council Role
GIO (Gemini) = The Researcher
- Market analysis & sentiment
- YouTube strategy extraction
- Pattern recognition
- 4-day research cycle driver
4-Day Research Cycle
┌─────────────────────────────────────────────────────────┐
│ DAY 1-2: RESEARCH │
│ ├── GIO searches YouTube for trading strategies │
│ ├── Extracts key concepts, indicators, rules │
│ └── Stores findings in vector_store │
├─────────────────────────────────────────────────────────┤
│ DAY 3: IMPLEMENT │
│ ├── ArchitectForge builds strategies from research │
│ ├── Deploys to paper trading environment │
│ └── Sets up performance tracking │
├─────────────────────────────────────────────────────────┤
│ DAY 4: TEST │
│ ├── Paper trades execute │
│ ├── Performance metrics collected │
│ └── Win rate, profit factor calculated │
├─────────────────────────────────────────────────────────┤
│ DAY 5+: PURGE & PROMOTE │
│ ├── Underperformers (<50% win rate) deleted │
│ ├── Winners promoted to live consideration │
│ └── Cycle repeats with new research │
└─────────────────────────────────────────────────────────┘
Fortress Integration
| Component | Fortress Layer | Purpose |
|---|---|---|
| gemini_agent | Layer 3 (AI Agents) | Research & analysis |
| strategy_extractor | Layer 3 (AI Agents) | YouTube → strategy |
| sentiment_tracker | Layer 5 (Observability) | Market mood |
| pattern_detector | Layer 3 (AI Agents) | Chart patterns |
Key Modules
neural_hub/
├── backend/
│ ├── gemini_agent.py # GIO API integration
│ ├── youtube_scanner.py # Strategy extraction
│ └── sentiment.py # Social sentiment
├── research_cycle/
│ ├── orchestrator.py # 4-day cycle manager
│ └── performance.py # Strategy scoring
└── api.py # FastAPI backend (port 8000)
Research Cycle Commands
# Start research phase
python -m neural_hub.research_cycle.orchestrator --phase research
# Check strategy performance
python -m neural_hub.research_cycle.performance --list
# Purge underperformers
python -m neural_hub.research_cycle.orchestrator --phase purge --threshold 0.5
YouTube Strategy Extraction
# GIO extracts strategies from YouTube videos
from neural_hub.backend.youtube_scanner import extract_strategy
strategy = extract_strategy(
video_url="https://youtube.com/watch?v=...",
focus=["entry_rules", "exit_rules", "indicators"]
)
# Returns structured strategy for ArchitectForge
Status
- Files: 6+ AI modules
- GIO: Active (Gemini API)
- Research Cycle: Defined, implementing
- Port: 8000 (FastAPI backend)