GoldNexa Data Sources & Methodology

This page explains, step by step, how the rates you see on GoldNexa are built: which third-party providers we use, how often data refreshes, the exact conversion formulas, and the limits of what an aggregated reference rate can tell you.

Last updated: August 2026

The short version

GoldNexa does not receive prices from any jeweller, exchange or bank. We start from international spot prices, convert them into Indian rupees per gram, and adjust for India's import duty so the figure lines up with published Indian reference rates. Everything shown is an indicative market-reference rate — a benchmark for comparison, not a quote you can transact on.

How the India rate is calculated

International markets quote gold in US dollars per troy ounce. Here is the full chain from that quote to the number on a GoldNexa rate page:

The final number is what we call the GoldNexa indicative reference rate.

Data providers

Content feeds: crypto and currency

Three additional free, key-less feeds power our market sections, all fetched and cached server-side so the upstream APIs are never hammered:

• Crypto prices — CoinGecko public API (top 50 coins by market cap, in INR), refreshed every 10 minutes.
• Currency rates — Frankfurter.app, serving the European Central Bank’s daily reference rates for 30 currencies, plus a one-year USD/INR history. The live USD/INR figure uses our own hourly market feed.

Update frequency, honestly stated

• Rates pages refresh every hour. Gold, silver and forex are polled hourly and platinum every 8 hours; updates are pushed to open pages automatically.
• This is not a real-time trading feed. During the hour between refreshes the displayed rate can lag a fast-moving market. Quoted intraday moves may not yet be reflected.
• Every rate page shows the timestamp of the data it is displaying, taken from the feed itself. If the feed is unreachable, the page says the data is temporarily unavailable rather than showing a made-up number.

What is and isn't inside the rate

City rates

The city selector adjusts the national reference rate by a small indicative markup per city. These markups are estimates of typical local premium differences — they are NOT a live feed from any city's jewellers or association. Two neighbouring shops in the same city can quote different rates. Use the city figure to compare ballpark levels between cities over time, not to predict an exact bill.

Historical chart data

Chart series are built from GoldNexa's own stored daily closes, which are derived from the same feeds above (each historical day uses that day's actual USD/INR rate). We never invent a price to make a chart look smooth — every point traces back to a real market close. Two things are worth stating precisely, because “no invented data” is easy to claim and easy to get wrong:

• Gap backfill uses real market closes, not guesses. Our host sleeps when idle, so days can be missed. We fill those from Yahoo Finance daily closes (gold GC=F, silver SI=F, platinum PL=F, palladium PA=F), each converted at that day's actual USD/INR rate. Backfill only ever fills a missing date — it never overwrites a day we recorded live.
• 5 points in the 10-year seed were interpolated. Of 2,513 trading days in the 2016–2026 seed, five isolated forex/palladium values were missing from the source and were filled by linear interpolation from the nearest surrounding days. That is 0.2% of the dataset, and the charts page discloses it too.

Days we could not fill from any real source appear as gaps, not invented points.

API call budget (how we never run dry)

Every rate-limited free tier is budgeted so the site keeps updating every single day of the month, with wide safety margins:

When a quota or network hiccup occurs, caches keep serving the last good values with their timestamp — pages never go blank and never show invented numbers.

Known limitations

• Aggregated reference rates will not match every local jeweller to the rupee — dealer premiums, making charges and local demand all vary.
• Between hourly refreshes, prices can move in the market before they move on GoldNexa.
• Cross-checks reduce bad readings but no free data feed is flawless; a rare stale or glitched upstream value can slip through.
• MCX Gold, where shown, is an estimate derived from spot (marked “est.”), not an exchange feed.

Spotted a problem?

If a rate looks wrong, tell us via the contact page — corrections from readers are the most useful feedback we get.