Every component of CryptoAI is designed with transparency and safety in mind.
From market data to execution, every step is visible and controllable.
The bridge between your LLM and the trading interface.
Trading strategies are expressed in natural language rather than code. Market data and system rules are supplied to the LLM as structured context, while final decision-making, validation, and execution remain fully controlled by the engine.
TradingEngine prepares structured JSON data payloads from market data and internal state, stored in SQL databases. LLMs operate on this data and respond in plain text, leaving execution logic fully under engine contro
The system applies deterministic validation rules to every action proposal. Only actions that comply with configured risk limits, exchange rules, and internal execution logic are forwarded to the exchange for execution.
Multiple layers of safety to keep you in control.
Execution can be paused or resumed at any time. Users can manually cancel pending orders or close open positions through the interface, independent of automated decision flow.
Current risk controls are applied at the strategy definition level via user-defined instructions. A separate execution-level circuit breaker system is under development to provide deterministic enforcement beyond prompt-based constraints.
All system activity is fully logged with precise timestamps. This includes market data snapshots, LLM decision outputs, execution attempts, order results, and manual user actions. Logs are stored locally and can be reviewed for analysis, debugging, and accountability.
Strategies can be evaluated using exchange testnets (e.g. Binance Testnet) or by disabling the execution layer to observe decision output without market interaction.
More integrations coming based on user demand.
Connect your preferred AI model for market analysis and decision making.
Join the early access program and shape the future of AI-assisted trading.
Request Early Access