How olaxbt/ai-market-maker works
AI Market Maker is a crypto hedge-fund-style trading stack with a web dashboard, paper-trading runtime, and research backtester. In the principal trading workflow, market data feeds specialist desks; their signals go through debate/arbitration and portfolio proposal, then Risk Guard can veto before execution. Separate dashboard/API surfaces support paper operations, backtests, and published results. The graph reflects README-described flow and source-index references; several subsystems are represented by their known entry modules because their implementations were not sampled.
Main entry points: main.py Trading workflow · engine.py Backtest engine · backtest_routes.py Backtest API · nexus_adapter.py Execution adapter
Components 23
Product surfaces
- Web dashboard
web/src/app/console/page.tsx - Trading API
src/api/main.py - Paper controls
src/api/engine_routes.py - Backtest API
src/api/backtest_routes.py - Results publishing
src/api/leadpage_routes.py
Trading workflow
- Trading workflow
src/main.py - Specialist desks
src/agents/technical_ta_engine.py - Desk debate
src/workflow/desk_debate.py - Signal arbitration
src/workflow/weighted_arbitrator.py - Portfolio proposal
src/agents/portfolio_management.py - Risk Guard
src/agents/governance/risk_guard.py
Market and configuration
- Market data providers
src/nexus_data/provider.py - LLM inference
src/llm/agent_llm_client.py - Policy and settings
src/config/app_settings.py
Backtest and results
- Backtest engine
src/backtest/engine.py - Perpetual simulator
src/backtest/engines/perp.py - Historical data
src/backtest/loaders/registry.py
Execution and records
- Execution adapter
src/adapters/nexus_adapter.py - Order management
src/oms/oms.py - Paper account
src/paper_account.py - Trace and event log
src/flow_log.py - Leaderboard storage
src/storage/leadpage_db.py
Other components
- Trader / operator
Connections 24
- Trader / operator → Web dashboard: uses
README.md - Web dashboard → Trading API: requests
- Trader / operator → Trading workflow: runs
README.md - Market data providers → Trading workflow: feeds market data
README.md - Trading workflow → Specialist desks: runs desks
src/main.py - Trading workflow → Desk debate: coordinates debate
README.md - Trading workflow → Signal arbitration: invokes arbitration
README.md - Trading workflow → Portfolio proposal: builds proposal
src/main.py - Trading workflow → Risk Guard: checks risk
README.md - Trading workflow → Execution adapter: executes approved trade
README.md - Specialist desks → LLM inference: uses inference
- Policy and settings → Trading workflow: configures
src/main.py - Trader / operator → Paper controls: controls paper loop
src/api/engine_routes.py - Paper controls → Paper account: reads and resets
src/api/engine_routes.py - Execution adapter → Paper account: uses paper state
src/adapters/nexus_adapter.py - Execution adapter → Order management: routes optionally
src/adapters/nexus_adapter.py - Trading workflow → Trace and event log: records events
src/flow_log.py - Trader / operator → Backtest API: starts research run
README.md - Backtest API → Backtest engine: runs backtest
src/api/backtest_routes.py - Backtest engine → Perpetual simulator: delegates simulation
src/backtest/engine.py - Backtest engine → Historical data: loads bars
- Backtest engine → Trading workflow: invokes workflow
src/backtest/engine.py - Backtest API → Results publishing: publishes results
- Results publishing → Leaderboard storage: reads and writes
src/api/leadpage_routes.py