Overview

For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to page URLs; this page is available as Markdown.

AI Chart Analysis — Overview

What AI Chart Analysis Is

Describe what you are analyzing, and watch the chart build the analysis.

AI Chart Analysis lets you describe what you want examined on a chart — in plain English — and watch the analysis appear directly on the price action: structure, support and resistance, Fibonacci levels, momentum, volatility, and more, drawn with exact coordinates and backed by named, checkable criteria.

You might type:

Analyze this breakout

Map the current market structure

Show the latest pullback

Draw the relevant Fibonacci retracement

Mark key support and resistance

Show momentum and the risk map

Analyze this chart

The feature exists to help you perform chart analysis faster, without first having to build a complete strategy or run a backtest.

How It Works

Behind the plain-English box, three things happen in sequence:

1. Your intent becomes a structured plan. An AI planner reads your prompt, your chart selection, and your preferences, and converts them into a structured, versioned analysis plan — not free-form text.

2. A deterministic engine executes the plan. It reads the OHLCV data, calculates indicators, identifies structures, evaluates conditions against named criteria, and works out exact drawing coordinates. The AI never invents a number, a level, or a coordinate — the engine is the only source of truth for anything that appears on your chart.

3. The chart is the output. Drawings, indicators, and structure markers appear directly on the price action, built up step by step so you can see how the analysis was assembled. A short text summary appears alongside the chart — the chart itself carries most of the information, so the text stays brief.

The full mechanics of that pipeline — planning rules, the deterministic engine's responsibilities, and the JSON contracts that connect them — are covered in Architecture and Data Contracts.

What Analysis Mode Is Not

AI Chart Analysis is an analysis and visualization tool — not a strategy generator, signal service, or execution system. Specifically, it does not:

• Generate a complete trading strategy

• Backtest or optimize a strategy

• Produce entry and exit signals as instructions to act on

• Execute trades or manage positions

• Guarantee or imply profitability

• Emit numeric probabilities of outcomes

• Replace the strategy builder

• Apply every available indicator to every chart

• Produce long-form chart commentary

• Depend on free-form AI-generated coordinates

• Repaint a drawing once it is confirmed

• Modify your chart permanently without your consent

If you already have a complete trading idea you want to turn into a managed, live-running algorithm, that's a separate feature: see Text-to-Algo (T2A). AI Chart Analysis and T2A intentionally don't overlap — Analysis Mode helps you see the chart more clearly; T2A helps you act on a fully-specified idea.

Compliance and Disclaimers

Because objects like target zones, risk maps, and invalidation levels can resemble investment recommendations in some jurisdictions, AI Chart Analysis is built around a strict compliance posture:

• Everything the feature draws or writes is presented as technical analysis of historical and current price data — never as advice to buy or sell.

• A standard disclaimer appears the first time you use Analysis Mode and remains accessible from every analysis summary afterward.

• Words like "buy," "sell," "enter," "exit," and "you should" never appear in generated summaries or labels. This restriction is enforced by the analysis engine and the summary templates themselves, not left to chance.

By using AI Chart Analysis, you acknowledge that its output is technical analysis of price data, not investment advice, and is not a guarantee of future performance.

What's Covered Elsewhere

Page	Covers
User Flow	The step-by-step flow from opening Analysis Mode to refining and saving a result
Analysis Types	Recognized keywords, supported analysis types, and automatic selection
Preferences	Persistent settings, learned preferences, and detail levels
Architecture	How planning and calculation are split between the AI and the deterministic engine
Data Contracts	The JSON contracts behind every plan and every engine result
Chart Behavior	What appears on the chart, and the no-repaint drawing state machine
Editing & Reuse	Refining, editing, saving, and rerunning an analysis
Reliability & Limits	Edge-case handling, latency budgets, and success criteria

For the exchange's other AI surfaces, see AI Copilot (the in-app conversational assistant) and Text-to-Algo (T2A) (natural-language strategy creation).