AlgoTradeALGOTRADEKIT PRODUCT GUIDE
AI-assisted trading with AlgoTradeKit
Understand how AlgoTradeKit uses analysis and reviewer models, verification and account permissions for AI-assisted trading.
Research is the starting point
AlgoTradeKit brings technical market context and configured AI models into a single workspace. An assessment can highlight a possible setup, explain uncertainty or recommend waiting. A directional signal is not an order and is not a probability of success. The quality of an assessment depends on the available evidence, data freshness and model capabilities.
Two perspectives before a proposal
The analysis model evaluates available evidence. An independent reviewer challenges the proposal and identifies missing context or conflicting signals. Model agreement alone does not establish profitability. Local models through Ollama, hosted OpenAI connections and hybrid routing are configuration options; hosted usage may incur provider charges. AlgoTradeKit is an independent product and is not endorsed by OpenAI.
Verification and execution are separate
Saving an AI connection does not grant order permission. Model access verification, evaluation and execution promotion are distinct steps. Passing an evaluation checks specified behaviors; it does not prove future trading performance. Account readiness, broker authorization, costs, market evidence and configured risk limits still apply.
What you control
Review the active broker, risk limits and Auto Trade permission in your account. Signing in or opening a chart does not start trading. Research-only mode allows review without autonomous order proposals being authorized for execution. Protective monitoring depends on the backend worker running; pausing new entries is not the same as closing an existing position.
Before enabling automation
Check the freshness of the data, the active account and its permissions, the model status and the applicable costs. Review how existing positions will be monitored. Markets can move against a position and execution may be delayed or unavailable. No AI model, stop reference or risk rule can guarantee a positive result.