AI Stock Analysis

Measurement and interpretation are two different jobs

RSI, MACD and a moving average stack are arithmetic. The same candles always produce the same numbers, and no model is needed to calculate them. The genuinely hard part of technical analysis is holding twenty readings in your head at once and deciding which one outranks the others when they disagree.

That split is the whole design of the tool. Everything measurable is measured deterministically on the server, and the model is handed a finished measurement summary rather than a raw price series. It never invents an indicator value; it explains the ones it was given.

The chain from candle to conclusion

A request starts by pulling candle data from the public Binance and Yahoo Finance endpoints for the symbol and timeframe you chose. The server then computes the indicator set — RSI, MACD line, signal and histogram, the EMA 9/21/50/200 stack, Bollinger Bands, ATR, ADX, Stochastic, OBV, VWAP and volume ratio among them.

Two more passes run before the model sees anything: candlestick and chart patterns are detected, and support and resistance zones are derived by clustering pivot highs and lows in the price data. Only that packaged summary goes out for interpretation.

What comes back in the output panel

The response is structured rather than a paragraph of prose. You get an overall assessment, the signals that support it, the signals that argue against it, a risk level, and suggested entry, stop-loss and target levels with the resulting risk/reward ratio.

The last field matters most: the analysis states what would invalidate the scenario. A read that cannot be wrong is not a read, and naming the level that breaks the argument is what makes the rest of it checkable.

Where a machine read differs from your own

Reading a chart by hand is selective by nature. You notice the indicator that confirms what you already suspect and skim past the two that do not, and the further a chart is from your usual instruments the more you skip. A model working from a fixed measurement summary covers the same ground on every asset, in the same order, whether you find the setup interesting or not.

It also has real blind spots. It does not know your position size, your holding period, your tax situation or the news that broke ten minutes ago. Treat the output as a second opinion on the chart, not as a decision.

Markets and timeframes

Borsa Istanbul shares, US stocks and crypto pairs run through the same interface, and the market is detected from the symbol you type, so there is no exchange to select first. Five timeframes are available: 15 minutes, 1 hour, 4 hours, 1 day and 1 week.

The timeframe is not a cosmetic choice. The same symbol can read as overbought on a 15-minute chart and as an early trend on the weekly, and picking the one that matches your intended holding period does more for the quality of the answer than any single indicator.

One asset at a time, or fifty at once

Analysis is deliberately narrow — one symbol, one timeframe, in depth. When the question is broader, the Opportunity Radar sweeps 50 assets with the same local indicator maths and scores each one from 0 to 100, so you can pick which few deserve a full analysis.

Free plan and Pro

The free plan gives you 3 AI analyses and 1 radar scan per day with no sign-up. An allowance is only consumed once an analysis has completed successfully — a repeated request served from cache or a call that fails does not cost you anything.

Pro is 9.99 USD per month and adds unlimited analyses, 20 radar scans a day, a visual read of the chart image by the model, analysis history and PNG export. Payment runs through Dodo Payments and can be cancelled at any time.