FOR SYSTEMS AND QUANT TRADERS
GOLD DATA FOR SYSTEMS TRADERS
Everything Gold Hunts publishes is available as JSON or CSV, documented in an OpenAPI 3.1 spec and exposed as MCP tools. Copy a sample below and you have gold data in a dataframe inside a minute.
Gold Hunts publishes gold data as stable JSON endpoints and CSV datasets: session signals with entry, stop and target, the broker-verified XAUUSD win-rate record, month-end gold prices aligned to five macro series, and an hourly volatility profile. An OpenAPI 3.1 spec and MCP tool catalogue make it callable from code and agents.
Endpoints
Every endpoint is a static JSON document served from the same origin as the site, which means no key, no rate limit, no auth handshake and no SDK. They are regenerated on a schedule and each payload carries its own generated_at timestamp and licence line, so a cached copy can always be dated.
Design decisions worth knowing before you build on them. Schemas are additive: fields get added, existing field names and types do not change. Numbers are numbers, not strings. Timestamps are ISO 8601 in UTC. Nulls appear where a value is genuinely unknown rather than being filled with a zero, because a zero in a gold price series is worse than a gap.
| Endpoint | Contents | Refresh |
|---|---|---|
| /api/gold-signals.json | Published session calls: direction, entry, stop, target, session, conviction, outcome | Hourly |
| /api/gold-record.json | Verified XAUUSD win rates and sample sizes by session and direction | Hourly |
| /api/gold-sessions.json | Asian, London and New York session windows in UTC | Static |
| /api/gold-glossary.json | Gold terminology with definitions | On change |
| /api/calculators.json | Formula specification for all 21 gold calculators | On change |
| /api/gold-research.json | Study index with method and sample metadata | On change |
| /openapi.json | OpenAPI 3.1 specification for every endpoint above | On change |
| /mcp-tools.json | MCP tool catalogue for agent frameworks | On change |
No P&L in the API
The record endpoint exposes win rates and sample sizes only. There is no field for profit, return or account balance, and none will be added. Build expectancy models on your own fills, not on this record.
Code samples
Three ways in: pull the record and pivot the session table in Python, poll the signal feed in TypeScript, or check the current call from the shell. All three hit the same documents a browser does, so anything you see on the site is reproducible from code.
import pandas as pd
import requests
BASE = "https://goldhunts.com"
# Verified XAUUSD record: win rates and sample sizes, no P&L.
record = requests.get(f"{BASE}/api/gold-record.json", timeout=20).json()
print(record["total_trades"], record["win_rate"], record["sample_period"]["start"])
sessions = pd.DataFrame(record["by_session"])
pivot = sessions.pivot(index="session", columns="direction", values="win_rate")
print(pivot)
# Month-end gold close aligned to the 10-year TIPS real yield.
macro = pd.read_csv(f"{BASE}/data/gold-vs-real-yield-monthly.csv", parse_dates=["month"])
print(macro.corr(numeric_only=True))
# Average XAUUSD range by hour of the UTC day.
hourly = pd.read_csv(f"{BASE}/data/xauusd-hourly-volatility.csv")
print(hourly.sort_values(hourly.columns[1], ascending=False).head(5))Backtest datasets
The CSVs are the same files the site charts from, which is the point: if a chart on Gold Hunts shows a relationship, you can load the identical series and check it. All are month-end aligned unless stated, with a single header row, ISO dates and no thousands separators.
A caution on the monthly series. 140-odd month-end observations is enough to describe a relationship and nowhere near enough to trade one. Treat these as context datasets for regime work, not as a signal source. The hourly volatility profile and session range shares are the files with direct execution value, because they tell you when gold actually moves.
| Dataset | Path | Granularity |
|---|---|---|
| Month-end gold close | /data/gold-monthly-close.csv | Monthly |
| Gold vs 10y real yield | /data/gold-vs-real-yield-monthly.csv | Monthly |
| Gold vs broad dollar index | /data/gold-vs-dollar-index-monthly.csv | Monthly |
| Gold vs CPI index | /data/gold-vs-cpi-index-monthly.csv | Monthly |
| Gold vs breakeven inflation | /data/gold-vs-breakeven-inflation-monthly.csv | Monthly |
| Gold vs fed funds | /data/gold-vs-fed-funds-monthly.csv | Monthly |
| Gold/silver ratio | /data/gold-silver-ratio-monthly.csv | Monthly |
| Gold seasonality by calendar month | /data/gold-seasonality-monthly.csv | Aggregate |
| XAUUSD hourly volatility profile | /data/xauusd-hourly-volatility.csv | Hour of UTC day |
| Session share of daily range | /data/gold-session-range-share.csv | Aggregate |
| Gold price vs miner margin | /data/gold-price-vs-miner-margin.csv | Monthly |
Contract mechanics your code needs to get right
Most gold position-sizing bugs come from three places: the contract size, the symbol name, and the assumption that a pip on gold behaves like a pip on a currency pair. A standard gold CFD lot is 100 troy ounces, so a $1.00 move in XAUUSD is $100 per lot and a $0.01 move is $1.00. Size from the dollar distance to your stop, never from a pip count.
Symbol names are broker-specific: XAUUSD, GOLD, XAUUSD.r, XAUUSD_ and worse all exist, sometimes with a suffix that changes between account types on the same broker. Resolve the symbol at runtime from the terminal rather than hardcoding it, and assert the contract size from the symbol specification before sizing anything.
Spread and swap on gold are materially wider than on major FX and vary hard by session. A backtest run on mid-price with a fixed 20-cent spread will look better than anything you can execute during the Asian session. The calculator specification at /api/calculators.json publishes the exact formulae the site uses for spread cost, swap, margin and position size, so your implementation can be checked against a reference rather than reinvented.
- Standard lot = 100 troy ounces; $1.00 price move = $100 per lot.
- Resolve the gold symbol from the terminal at runtime; never hardcode a broker suffix.
- Assert contract size from the symbol spec before computing position size.
- Model spread per session, not as a constant.
- Include swap on any position held over the broker rollover.
Reference pages
- GOLD DATA API
Full endpoint documentation, schemas and licence text.
- GOLD SYMBOL NAMES ON MT5
How brokers name gold and why runtime resolution matters.
- GOLD CALCULATORS
Reference formulae, all published at /api/calculators.json.
- TESTING GOLD STRATEGIES
Sample size, decay and demo-to-live on a volatile metal.
Licence and fair use
Free to use, including commercially, with visible attribution to Gold Hunts. That covers pulling the endpoints into a research stack, redistributing derived figures, and embedding the charts. Polling politely is the only real constraint: the documents are static and cached, so hourly is generous and per-second adds nothing but load.
If your system needs guarantees the free layer cannot make — an SLA, a change-notification channel, deeper history or higher cadence — email info@goldhunts.com. Better to have that conversation before a production system depends on a static file.
GOLD TRADING FAQ
- Does the Gold Hunts API need an API key?
- No. Every endpoint is a static JSON document served over HTTPS with no key, no auth and no rate limit. Poll hourly at most, since the documents regenerate on a schedule.
- Can I use Gold Hunts data in a commercial trading system?
- Yes, with visible attribution to Gold Hunts. For an SLA, change notifications, deeper history or higher refresh frequency, contact the desk to arrange it rather than relying on the free static layer.
- Which Gold Hunts dataset is most useful for execution?
- The hourly XAUUSD volatility profile and the session range shares. They tell you which hours of the UTC day gold actually moves in, which drives session selection and stop placement. The monthly macro series are context, not signals.
- How large is a standard gold lot for position sizing?
- 100 troy ounces, so a $1.00 move in XAUUSD equals $100 per standard lot. Size from the dollar distance between entry and stop rather than from a pip count, and confirm the contract size from the broker symbol specification at runtime.
- Is there an MCP server for Gold Hunts?
- Yes. The tool catalogue is published at /mcp-tools.json alongside an OpenAPI 3.1 spec at /openapi.json, so agents can call the gold record, signal feed, session windows and calculator formulae directly.
TRACK YOUR OWN GOLD RESULTS FREE
Live XAUUSD signals on a public ledger, a free gold trade journal that syncs with MetaTrader 5, and the macro pulse before the London open.
RELATED ON GOLDHUNTS
- Gold statistics for journalists — citable data, charts and datasetsCluster hub
- Gold day trading workflow — a repeatable XAUUSD routineSame cluster
- Gold research for funds — methodology, verified record and data licensingSame cluster
- How the gold price is actually set — LBMA fix, spot and physicalSame cluster
- Free Gold Trading CalculatorsCalculator
- Gold Hunts Alternatives for Gold TradersComparison