Validate

Check the trade before you take it.

Name: Versatum
Availability: OnlineOnly
Author: Versatum

AI gates plus chart math you can audit. A 16-stage Trade Validator scores your idea across 12 weighted dimensions and 14 gating rules; Chart Signals checks the chart with three deterministic studies and zero model calls; a vision pass reads 14 classical chart patterns; and Watchlist Health and the AI Risk Mapper put the names you already follow through the same scrutiny.

Start Free Then place it

Trade Validator

AI validates every trade.

Before you enter a position, the Trade Validator runs a 16-stage pipeline in five waves and scores the trade across 12 weighted dimensions — technical, fundamental, sentiment, momentum, risk and more — against 14 gating rules.

A single head-of-desk synthesis call turns that into one of five verdicts, from STRONG_BUY to STRONG_SELL, with risk-adjusted sizing. High-confluence setups get the green light. Low-conviction trades get flagged with specific concerns — so you never enter a position blind. Every recommendation is explained: the confluence score is broken down by weighted dimension, and every gating rule it applied is named.

Mode-aware confluence scoring — 12 weighted dimensions, 14 gating rules, five verdict bands
16 stages in 5 waves; exactly one of them calls a model — the head-of-desk synthesis at the end
Five verdicts from STRONG_BUY to STRONG_SELL, each with risk-adjusted sizing
Every recommendation explained: the confluence score broken down by weighted dimension, and every gating rule it applied named
Stocks and crypto — on crypto it adds BTC correlation scoring (the AI Trade Validator tab of the 11-tab crypto desk); it is not on the options or FX desks
Morning Picks runs the same validator pass on its top 70 names before the bell
Free plan: 3 runs a day; Pro and Pro+ run it with no daily cap

The same validator is a stage inside Morning Picks, where the top seventy names fan out to a validator pass before up to twenty setups publish. Once it has scored your idea, check whether the chart agrees with Chart Signals and Chart Patterns, then place it on paper or live.

Trade Validator

MSFT|BUY

MSFT

BUY

Confluence Score

Low RiskHigh

TechnicalRSI 42 | above 20 EMA | MACD bullish cross83%

FundamentalDCF $485 vs market $412 | margins expanding91%

Sentiment28 analysts: Strong Buy consensus67%

MomentumTrend intact | volume rising | RS 7276%

RiskVIX elevated at 18.2 | sector rotation risk58%

AI Reasoning

RSI at 42 — not overbought, room to run on daily timeframe

DCF fair value $485 vs current $412 — 18% margin of safety

Analyst consensus: Strong Buy (28/34 analysts) — recent upgrades

5-dimension confluence|AI-validated trade scoring

1/4

Trade Scoring

Confluence scoring.

Trade Validator runs 16 stages in 5 waves and blends 12 weighted dimensions into one score, then applies 14 gating rules and lands it in one of five verdict bands. Exactly one of those stages calls a model — the head-of-desk synthesis at the end. Everything feeding it is arithmetic you can check.

Weights are per trading style: day traders get momentum, VWAP and intraday flow; position traders get fundamentals and analyst targets. A dimension we could not measure for a ticker is dropped from the denominator rather than scored zero, so a thinly covered name is not quietly punished for our missing data.

Swing-trade weights · sums to 100

Chart Pattern

Momentum & Technicals

Sector / Industry Rotation

News Sentiment

Analyst Price Target

Market Regime

Fundamentals & Valuation

Gap Analysis

The one page where we chose not to use AI.

Does Versatum have indicators that aren't AI? Yes. Chart Signals is the one research surface here with no model in it at all: three studies, written line by line from their reference source into TypeScript and run on our own bars. Every number is deterministic, reproducible and pinned to a committed source hash — same inputs, same bars, same numbers, every time: nothing is inferred, summarised or generated. Six bar intervals, seven lookbacks, and it works on stocks, crypto and forex.

No AI · Fully Deterministic

The confirmation step, with no model in it.

Chart Signals is the one desk here that no model touches — the one workspace where no model has an opinion. Run it after the Validator has scored your idea, to see whether the chart is confirming it: overlay a study on stocks, crypto or forex across six bar intervals and seven lookbacks — no tier gate, no daily cap — and read the levels, bands and tags straight off the price chart.

Each study carries its provenance on screen, printed beside the chart: the script revision it was ported from, the twelve-character (12-hex) hash of the committed reference source it is pinned to, how many bars it actually ran on, the interval those bars were delivered at, and every one of the 26 places the port deliberately departs from the reference implementation. Its 55 inputs are split into three classes — compute, window and presentation — and each control is badged with which it is and what it moves, so changing what you are looking at provably cannot change what the study read. The bar producer declares the interval it delivered, which means a study refuses to compute on bars of an unstated interval instead of guessing. 1,637 contract checks and a 379-mutation sweep stand behind all of it.

3 studies · zero AI

55 inputs in 3 classes

6 intervals · 7 lookbacks

Pinned to a committed source hash

1,637 checks · 379 mutations

Three studies: Follow-Through Levels, Money-Flow Trail, Indicator Consensus — exactly three, not a library
Zero LLM calls — same inputs, same numbers, every time; deterministic script code, not a model
Deterministic calculations, validated against a committed source hash
A TypeScript port of each Pine script, hand-transliterated line by line from its reference source
6 bar intervals (5m, 15m, 30m, 1h, 4h, Daily) × 7 lookbacks (Same Day → 1 Year)
Works on stocks, crypto, and forex — the same three studies on equity, crypto and FX bars
Inputs split into compute / window / presentation, so changing the view cannot change a reading
Refuses to compute on bars of an unstated interval instead of guessing
1,637 contract checks behind the pipeline, plus a 379-mutation sweep
No tier gate and no daily cap — on every plan, Free included

Chart Signals

No model in the loop

Follow-Through Levels4 inputs

Historical-frequency levels above the previous high and below the previous low, each with a right-hand probability reading, fill bands, and a WIN / LOSS / Profitability panel.

Money-Flow Trail11 inputs

A per-bar trailing stop-and-reverse line, its two-bar-lag twin, a colour-flipping band between them, and confirmed BUY / SELL tags.

Indicator Consensus40 inputs

A weighted scoreboard, not a network: 29 switchable contributions, 10 of them paying at defaults, over a five-EMA ribbon and a Hull average.

Printed beside every reading

Script revision · a 12-hex hash of the pinned reference source · the bar count · the interval actually delivered · all 26 declared deviations from source.

LLM calls

1,637

Contract checks

379

Mutations swept

Follow-Through Levels

4 inputs

Ports the published “Breakout Probability (Expo)” script

Up to ten historical-frequency price levels — up to five above the previous high and five below the previous low, 1% apart by default — each with its own right-hand probability reading (the share of past sessions that actually reached it), fill bands, and a WIN / LOSS / Profitability panel. Levels that measure 0.00% — the ones that never printed — are hidden, so the count is whatever the data supports rather than a fixed number.

Money-Flow Trail

11 inputs

Ports an “AlphaTrend” script

A per-bar trailing stop-and-reverse line, its two-bar-lag twin, and a band between them that flips colour with the trend. BUY and SELL tags print only once the cross — the flip — is confirmed.

Indicator Consensus

40 inputs

Ports a TradingView script published as “Neural Network”

Named after the published script it came from, and it is not one: it is a weighted scoreboard, not a neural network — no training, no weights fitted to data. 29 switchable contributions over 16 computed indicators — 10 of them paying at defaults — plus a five-EMA ribbon, a Hull average, and a live 52-cell on-screen panel naming every contribution currently paying, so you can see exactly what moved the score.

Zero model calls

Nothing here is sent to a model. The same bars always produce the same reading, which is not something an LLM can promise you.

Pinned to its source

Each study prints its script revision and a 12-hex hash of the committed reference file it was written from, next to the reading itself.

1,637 contract checks

Plus a 379-mutation sweep that deliberately breaks the maths to prove those checks would notice. All 26 places where we knowingly deviate from the original scripts are listed on screen.

The view cannot move the number

All 55 inputs are classed compute, window or presentation, and every control tells you which — so changing what you are looking at provably cannot change what it reads.

Auditable by design

One surface you can check line by line.

Most of security is about keeping other people out of your data. Chart Signals is about a different kind of trust: being able to verify what the software told you.

Three studies, written line by line from their reference source and run on our own bars. Each one ships beside the script it was ported from, pinned by a committed 12-character source hash, so the transliteration can be diffed against the original line by line. No model touches this surface, so the same bars always produce the same reading. Inputs are split into three classes — compute, window and presentation — which is what makes changing your view provably unable to change a measurement. And where our port departs from the source, the study says so on screen rather than in a changelog.

One clarification we'd rather make ourselves: the study we ship as Indicator Consensus ports a TradingView script published as "Neural Network". It is a weighted scoreboard, not a neural network — no training, no weights fitted to data. Its panel names every contribution that is currently paying, so you can see exactly what moved the score.

studies shipped

Follow-Through Levels · Money-Flow Trail · Indicator Consensus

model calls

deterministic and reproducible end to end

declared deviations

every one shown on screen

1,637

contract checks in CI

plus a 379-mutation sweep, all 379 killed

In the app you will find it under Validate, beside Trade Validator and Chart Patterns — the answer to "is the chart confirming my idea?" that you can audit yourself, with no tier gate and no daily cap.

Vision AI · Chart Patterns

Chart pattern detection.

A vision LLM scans the chart and surfaces the patterns a senior trader would. The vision model is handed a rendered candlestick chart and asked for 14 classical patterns, each with a confidence score, a completion percentage, a target, a stop-loss and the support and resistance it read.

A run is three model calls — two vision passes and one text pass — and any pattern below 50 out of 100 confidence is thrown away rather than shown to you. Ascending triangles, cup-and-handle, head-and-shoulders, breakout structures — detected with confidence scoring and price targets, across stocks, forex and crypto with asset-aware prompts. Inside the Trade Validator, Chart Pattern is the heaviest swing-trade dimension, at 18 of 100.

When a chart read fails, we say so. Morning Picks records the reason, shows a "Some analyses failed" banner with the count, and the synthesis prompt is explicitly told not to infer a chart it never saw — so a briefing built without chart input can no longer be mistaken for a healthy one. There, only the twenty names that clear the downselect get the vision model; it is the expensive step, so it never runs on all seventy.

Head & Shoulders

Double Top/Bottom

Triangles

Channels & Flags

Cup & Handle

Wedges

14 classical patterns — head & shoulders, double tops and bottoms, triangles, flags, wedges, channels, cup & handle, plus ascending triangles and breakout structures
Confidence, completion %, target, stop-loss and support / resistance on every pattern that survives
Anything under 50/100 confidence is discarded rather than shown
Three model calls per run: two vision passes and one text pass
Stocks, forex and crypto, with asset-aware prompts — it is the AI Chart Pattern tab on the 6-tab forex and 11-tab crypto desks
One vision model reads the chart images — a separate model from the text model
Chart images for the vision model are drawn in your browser; nothing is uploaded until you run the analysis
Free plan: 1 run a day, counted atomically and reset at UTC midnight; unlimited on Pro and Pro+

Chart Patterns

AAPL|AI Vision

Chart Render

Candlestick Chart

AAPL3M

Patterns Found

Head & Shoulders

bearish92% complete

87%

Rising Wedge

bearish78% complete

71%

Bull Engulfing

bullish100% complete

64%

Key Levels

Support

$188.5

$185.2

$180.0

Resistance

$196.8

$201.5

$205.0

Trade Signal

SELL

Confidence82%

Entry$192.40

Target$185.20

Stop$197.50

R:R1:1.4

H&S neckline break confirmed

Volume declining on right shoulder

Target = head-to-neckline distance

Vision AI|14 pattern types

1/4

Portfolio Checks

Check what you already hold, too.

Validation is not only for new ideas. Two Watchlist sub-tabs — Watchlist Health and the AI Risk Mapper — score each name you follow and map how the list moves together, and each gets a card of its own.

Analytics

Watchlist Health

A 0-100 health score per name on your watchlist, each with a hold / add / trim / remove verdict and the specific issues behind it.

AI health monitoring on persistent watchlists synced via Supabase with polled quotes — a Watchlist sub-tab beside the Risk Mapper.

Risk

AI Risk Mapper

Your watchlist as a correlation map: pairwise matrix, diversification score, sector mix and average beta vs SPY over the names you follow — each reported with how many of them it could be measured on, so its measurement coverage in names is on screen.

A Watchlist sub-tab beside Health; listed with the Pro plan, alongside the Economic Calendar.

Validate first.
Then place the trade.

A 16-stage Trade Validator, three deterministic Chart Signals studies with no model in them, a vision read of 14 classical chart patterns, and health and risk checks on your watchlist — every verdict explained, every chart reading reproducible.

Start Free See Pricing