Multi-asset AI trading system: equities + energy futures, governed by an AI swarm with MCP tools.

• Overview
• Architecture
• Color-coded repo map
• Diagrams & flow trees
• Setup (Phase 1: Skeleton)
• Build Phases
• Usage (Next Steps)
• Contributing
• License
• Contact

HydraGrid is a modular, multi-asset trading platform designed for:

• Equity/ETF trading (US + global)
• Energy market trading (power futures, gas/oil futures, spreads)
• AI-driven strategy orchestration (swarm of agents + MCP tools)
• Multi-tenant SaaS delivery (API + web portal)
• Extensibility (new venues, instruments, strategies without rewrites)

See ARCHITECTURE.md for the full design document.

hydragrid/
  packages/
    engine-core/           → Universal order/portfolio engine
    data-pipelines/        → Stock & energy data ingestion
    strategies-stock/      → Equity strategies
    strategies-energy/     → Futures/energy strategies
    swarm/                 → AI agents + MCP orchestration
    risk/                  → Risk engine, limits, scenarios
    analytics/             → Metrics, backtests, benchmarks
    infra/                 → API, workers, deployment
    web-portal/            → Frontend + user dashboard
    docs/                  → Documentation

Legend: packages are grouped by functional domain and color-coded consistently across docs and diagrams.

• Data:
• Engine:
• Strategies:
• Risk:
• Swarm (AI/MCP):
• Analytics:
• Infra/API:
• On-chain:

Package	Domain	Path
	Engine core	packages/engine-core/unykornx_engine/
	Data pipelines	packages/data-pipelines/hydragrid_data/
	Stock strategies	packages/strategies-stock/unykornx_strategies_stock/
	Energy strategies	packages/strategies-energy/unykornx_strategies_energy/
	AI agents + MCP	packages/swarm/hydragrid_swarm/
	Risk engine	packages/risk/hydragrid_risk/
	Analytics	packages/analytics/hydragrid_analytics/
	API, workers, deploy	packages/infra/hydragrid_infra/
	Contracts + scripts	packages/onchain/

Color legend and end-to-end diagrams are in DIAGRAMS.md.

See consolidated diagrams (system architecture, data flow, swarm decisioning, on-chain control plane, risk policy flows) in DIAGRAMS.md.

• Python 3.10+
• pip or pipenv

1. Clone the repo:

   cd hydragrid

2. Create a virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # or .venv\Scripts\activate on Windows

3. Install in development mode:

   pip install -e .[all]

4. Copy .env.example to .env and configure:

   cp .env.example .env

pytest tests -v

• Phase 1: Skeleton + VS Code workspace ✓
• Phase 2: Engine-core + toy strategies + backtest
• Phase 3: Analytics & benchmarks
• Phase 4: Swarm + MCP tools
• Phase 5: API & Portal
• Phase 6: Energy expansion + real venues

After Phase 1 scaffold is complete:

1. Point builder-AI at packages/engine-core/ to implement Phase 2.
2. Each phase unlocks the next with increasing complexity.

See PHASE_GUIDE.md for detailed implementation steps per phase.

Contributions follow the project structure and coding standards in .vscode/settings.json and pyproject.toml.

MIT. See LICENSE.

HydraGrid Dev Team