REGIMES AI — Gold Intelligence Terminal

This is AI research & analysis, not financial advice. Outputs are the model’s read of public data and may be wrong — any decision is your own.

today's gold read · XAU/USD

live · $4,330/oz

Verdict

Model bias: bullish · entry quality: low

live price $4,330 /oz

Confidence

Risk levelMedium-High

Model read: structurally supported, but short-term momentum is weak and entry quality is low.

Model invalidation: the model's bullish thesis weakens if gold breaks below $4,150.

AI forecast · multi-horizon ?

spot $4,330/oz

Tomorrow · 1D

NEUTRAL

$4,295–$4,375

median $4,335 · conf 54%

7 days

BULLISH

$4,300–$4,520

median $4,405 · conf 60%

30 days

BULLISH

$4,186–$4,933

median $4,516 · conf 58%

90 days

STRONG BULLISH

$4,400–$5,400

median $4,850 · conf 63%

Short-term is soft (post-drop), long-term stays bullish — horizons are allowed to disagree. Full forecast cone →

why this matters

Gold is still structurally supported, but short-term momentum is weak. The AI suggests waiting before buying aggressively because macro risk and event risk are still elevated.

what is the model signaling?

NEUTRAL · BIAS BULLISH

The model reads gold as structurally supported but with low short-term entry quality. This describes what the model sees in current data — it is information, not a recommendation. Any decision is your own.

model accumulation map · 30 days ?

illustrative

30-day target ?

—

Plan confidence ?

—

moderate · split the buy

$4,400

weight 30%

once price stabilizes above this

$4,200

weight 35%

on dips here, if support holds

$4,516

weight 35%

after CPI + FOMC clear (target zone)

These zones are derived from the model's distribution; the model's invalidation level is $4,150. Informational only, not advice.

› WHAT THE MODEL READS

Is the model's monthly read constructive?

Model: bias is constructive (structural), but it flags low near-term entry quality.

How does the model rate this zone?

Model: becoming interesting in its read, but it still wants confirmation.

Does the model read gold as overheated?

Model: not extremely overheated, but risk stays elevated before major macro events.

How does the model treat FOMC/CPI?

Model: it discounts breakouts around those prints; event risk can shift the thesis quickly.

what changed today?

1. Macro pressure increased.

2. Gold momentum weakened.

3. Safe-haven demand still supports the bigger trend.

4. Model verdict shifted from BUY to NEUTRAL.

watch next

CPI / FOMC US dollar strength Real yield movement Gold reclaiming $4,400

Want to see the full AI reasoning — 10 agents, indicators, track record?

This is AI research & analysis, not financial advice. Outputs are the model’s read of public data and may be wrong — any decision is your own.

XAU / USD — today's verdict ?

live

.00

▼ −3.9% week per oz · $139.21/g · $139,210/kg

Model bias: bullish (structural) · entry quality: low

model thesis ?

Loading the live model thesis…

Invalidation ?: weekly close below $4,150 or 10Y real yield above 2.4%.

primary drivers

Central-bank demand Geopolitical bid Secular uptrend

risk flags

Fed on hold CPI/FOMC <10d Momentum ↓

What changed today is shown live in the verdict and the “what changed today” panel.

confidence rate ?

trade-grade ≥ 60% · neutral < 55% · live

Brier score ?

—

No-trade risk ?

—

committee snapshot ?

Consensus

4·3·3

bull·neut·bear

Disagreement ?

moderate

Conviction ?

watch

key levels

Support$4,150

Resistance$4,400

30d target$4,516

Next checkpoint: US CPI in 3d · FOMC in 10d

thesis change alert ? · 2h ago

Gold thesis changed from BULLISH → WAIT / NEUTRAL because a strong jobs report reversed real yields higher and price lost $4,400 support. DXY firming adds pressure.

30d median ?

—

Heat composite ?

—

Bottom proximity ?

—

Trend strength ?

—

› LIVE CHART

XAU/USD candlestick ?

OANDA:XAUUSD

› FORECAST PRO · adjustable horizon

multi-horizon forecast ?

spot $4,330/oz

Tomorrow · 1D

NEUTRAL

$4,295–$4,375

median $4,335 · conf 54%

7 days

BULLISH

$4,300–$4,520

median $4,405 · conf 60%

30 days

BULLISH

$4,186–$4,933

median $4,516 · conf 58%

90 days

STRONG BULLISH

$4,400–$5,400

median $4,850 · conf 63%

adjustable horizon cone ?

Median target 30 days: $4,516 · range $4,186–$4,933

▬ median ▬ P25–P75 ▬ P10–P90

model output — probability distribution ?

Expected return

7D: —

30D: —

Committee lean

—

Model signal

—
from today's verdict

› ADVANCED INDICATORS AI synthesis

AI reads all indicators ?

A 20-indicator gold-specific suite (incl. real-yield fair-value gap, GVZ, gold/silver, macro-tailwind composite) feeds the agent committee’s weighted read — not just basic RSI/MACD.

Technical agent confidence: — · see the full agent grid below.

conviction

—

0–100 · live

› AI CANDLE INTELLIGENCE pattern engine

candle index ?

—

reading live candles…

scans 28 classic patterns on daily gold OHLC

PATTERN LIBRARY · 28 patterns · detected patterns light up with their status

Reliabilities are Bulkowski US-equity upper bounds — rare patterns down-weighted, invalidated signals contribute zero. Confirmed = price followed through past the trigger within 3 bars. Information, not advice.

› LIQUIDITY & CAPITAL money flow

global liquidity is gold’s tide ?

Tracking Fed net liquidity, money supply, reverse repo and financial conditions — the capital flows that move gold beyond price.

liquidity score

—

0–100 · live

› MULTI-AI-AGENT ENGINE 17 AI agents

LAST COMMITTEE SWEEP · — · committee idle between daily sweeps

17-agent swarm · scanning_

nodes drift = monitoring · pulse = sweep · lines = cross-checks

chief orchestrator ?

Primary reason

—

Counter-case ?

Fed on hold + sticky CPI keep real yields elevated, and momentum just broke lower.

Conflict resolution

— Risk Manager caps confidence at — (live).

Model signal

Model read: bias bullish, entry quality low · model invalidation $4,150.

debate scoreboard ?

Bull team

Bear team

Winning side—

Primary winning factor—

Judge note—

analyst swarm · 10 of 17 ? core = MVP

10 analysts below · + bull/bear/judge (debate) · + PM, calibration, memory, auditor (executive) = 17 total

agent reliability ranking · last 90 days ?

Agents below 50% are auto-muted until reliability recovers — that's how the engine stops trusting noise.

full committee · every agent ?

multi-model

8 analysts → regime gate → contrarian overlay → 3-agent debate → 5-agent executive. Each layer uses a different AI model for its strength.

› CONFIDENCE BREAKDOWN

why this number? ?

Final confidence—

no-trade / wait logic ?

no-trade risk

Reason: FOMC + CPI within 10 days, downside momentum, ATR elevated, confidence below trade-grade.

Model read: structurally bullish bias with low entry quality · model invalidation below $4,150.

confidence timeline · 30d ?

Confidence slid 71% → 58% over 30d as the macro turned — the WAIT call came before the price break.

what would change my mind? ?

The model turns bearish if any of these fire:

daily close below $4,150

10Y real yield above 2.4%

DXY breaks above 106

ETF outflows exceed 20t/week

Turns bullish again on a reclaim of $4,400 with real yields rolling over.

› TRACK RECORD

prediction accuracy ?

win rate · last 30 days

19/28 resolved directional calls · 2 pending

Live predictions ?

Wait/Neutral held ?

Sortino ?

2.1

Calmar ?

3.4

self-improving engine ?

model v3.142 · retrained nightly · last 02:00 UTC

Mistakes learned

142

Calibration err ?

0.31→0.18

Accuracy 12w ago

52%

Accuracy now

64%

Conf. stability ?

+22%

how it learns — the daily loop ?

01 · PREDICT

Issue verdict

10 agents → orchestrator → verdict + confidence, logged with timestamp.

02 · GRADE

Score at T+7d

Compare to the real move. Win, miss, or correctly-held. No hiding losers.

03 · LEARN

Mine the miss

Each wrong call → labelled example: which agent over/underweighted?

04 · RETRAIN

Re-fit nightly

Deep model + agent weights updated on new graded data at 02:00 UTC.

05 · IMPROVE ↺

Ship + repeat

New version goes live; calibration tightens; loop restarts.

model memory — what it learned ?

1. Real-yield jumps >+8bp/week hurt XAU more than any oversold-RSI bounce has helped.
2. Central-bank demand only cushions drawdowns when DXY is flat-to-down — not when DXY is rising.
3. CPI/FOMC weeks cut breakout reliability ~18%; pullback-buys beat breakout-buys by ~11%.
4. After 3%+ weekly drops, XAU stabilized within 5 sessions in 7 of the last 9 cases before resuming up.

last night's changes · v3.141 → v3.142

▲ Real-yield feature weight +6% — it led the last 3 correct calls.
▲ Down-weighted breakout signals during event weeks (lesson #3).
▲ Added a post-3%-drop stabilization prior (lesson #4).
▲ Tightened NEUTRAL band; fewer low-conviction false signals.

daily prediction log ?

Date	Verdict	Conf	XAU @ call	+7d	Change	Outcome
05 Jun	WAIT	58%	$4,505	$4,330	−3.9%	held
29 May	BUY	67%	$4,410	$4,505	+2.2%	✓ win
22 May	BUY	71%	$4,290	$4,410	+2.8%	✓ win
15 May	NEUTRAL	49%	$4,320	$4,305	−0.3%	held
08 May	SELL	63%	$4,480	$4,360	−2.7%	✓ win
01 May	BUY	69%	$4,250	$4,210	−0.9%	✗ miss

Live calls are logged daily and graded at T+7d. This table fills in as calls resolve.

miss autopsy ?

› REGIME DETECTOR

current regime ?

SECULAR BULL · TACTICAL PULLBACK

real-yield & geopolitics driven

Dominant drivers ?

1 · Real yield 2 · Fed path 3 · Central-bank demand 4 · Geopolitical risk

De-emphasized ?

1 · RSI 2 · Retail sentiment 3 · Short-term candle patterns

regime impact on agents ?

regime confidence —

Dollar & Yield

1.4×

up-weighted

Macro Regime

1.3×

up-weighted

Central Bank

1.2×

up-weighted

Technical

0.8×

down-weighted

Sentiment

0.6×

muted

Reason: historical performance in real-yield-driven regimes — yield/dollar agents lead, short-term technicals & sentiment add noise.

› TOOLS

live converter ?

amount

unit ?

currency

100 g gold =—

live spot + live FX · rates loading…

popular currencies · 100 g ?

AI analyst ?

regimes ai · analyst

Ask me anything about gold — today's verdict, the multi-horizon forecast, the macro regime, drivers, or what could move XAU next. I read the live committee data on this dashboard. XAU-gold questions only.

reminders & alerts ?

Thesis change alert ?armed

XAU stabilizes above $4,400price · armed

Invalidation $4,150risk · armed

US CPI releasein 3d · high

FOMC decisionin 10d · high

› FOR AI AGENTS (API)

hosted MCP ?

Let your agent pull XAU verdicts, thesis receipts, and confidence over a hosted MCP server — free API key.

tools: xau.daily_verdict · xau.action_plan · xau.calibration_report

# endpoint

https://xau.ai/mcp/
# header
X-API-Key: xau_••••

XAU — Alpha gold intelligence. Informational and research use only. Nothing here is investment, financial, legal or tax advice, an offer, or a recommendation to buy, sell or hold any asset. Outputs describe a model's reading of public data and may be wrong. You are solely responsible for your own decisions.
Live spot via gold-api.com · live FX via exchangerate-api · chart via TradingView · forecasts, agents & indicators are live output from the 17-agent committee engine (multi-model), refreshed daily.

FREE

PRO