Cracktrader Documentation

This docs site now treats the runtime architecture as the primary orientation layer.

If you are new to Cracktrader, or trying to re-ingest the repo after earlier Cerebro-centric docs, start with the architecture pages below before reading older concept pages.

Read This First

1. Architecture Index
2. Runtime Map
3. Mode Matrix
4. Runtime Terms

Current Runtime Model

The preferred mental model is:

• a session owns shared runtime services
• market inputs update a shared state coordinator
• the runtime builds one immutable snapshot per step
• strategies consume filtered views of that snapshot
• strategies emit attributed intents against execution contexts and routes
• inventory, risk, execution adapters, and post-trade hooks handle the rest

Backtrader compatibility is still important, but it is no longer the right top-level map for the system.

Where To Go Next

Goal	Start here
Understand the current architecture quickly	Architecture Index
See the runtime boundaries end to end	Runtime Map
Compare backtest, paper, sandbox, and live behavior	Mode Matrix
Look up stable runtime vocabulary	Runtime Terms
Understand which docs are canonical vs transitional	Docs Source of Truth and Migration Map
Follow legacy links or older mental models safely	Legacy Architecture Context
Install and run your first example	Getting Started
Find stable user-facing APIs	Reference

Runtime Overview

graph TD
    A["Session"] --> B["Reference Data"]
    A --> C["State Coordinator"]
    A --> D["Strategies"]
    A --> E["Execution Contexts and Routes"]
    A --> F["Inventory and Risk"]
    A --> G["Post-Trade and Control Plane"]
    H["Feeds and Channels"] --> C
    C --> I["Shared Snapshot"]
    I --> D
    D --> J["Execution Intents"]
    J --> E
    E --> F
    F --> G

Compatibility Note

Older pages about feeds, brokers, stores, and Cerebro are still available because they remain useful for implementation detail and migration work.

Use them as secondary detail, not as the primary runtime truth.