Real-Time / Minute-Level Data Coverage · Privora

A-share/HK minute-level data + exchange ETF real-time quotes — coverage, frequency, and freshness spelled out

Most platforms just say their data is comprehensive without saying how often it refreshes or how fresh it actually is right now. This page lists Privora's stock, exchange-traded ETF, and precious-metal assets one by one — content, coverage, refresh frequency, and freshness — including an honest line on what counts as minute-level and what doesn't.

Preview minute-level data — no signup Sign in / register to start

Preview of minute-level and real-time market data samples in the Privora marketplace

Coverage, frequency, freshness — no adjectives, just the table

The table below lists the actual coverage and refresh frequency of Privora's main assets, based on a single measurement taken at 2026-07-29 16:16 — numbers quoted verbatim, not rounded up. We also spell out which data qualifies as minute-level and which doesn't, including the most commonly misunderstood one: stock_day.

Table 1: Content / Coverage / Frequency / Freshness

The freshness figures below (e.g. "about 70 seconds behind") come from a single sampling run at 2026-07-29 16:16 — they illustrate the order of magnitude, not a real-time guarantee for every visit. Check the actual timestamp you see after signing in.

Asset	Content	Coverage	Frequency	Freshness (sampled 2026-07-29 16:16)
`stock_minutes`	Five-level bid/ask snapshot — last price + bid1-5/ask1-5 price & size + day high/low/prev close	A-share SH 2,463 + SZ 3,081 + HK 2,866 = 8,410 stocks	About every 60-90 seconds intraday (Tencent 00700: 374 snapshots today, 09:00:23 -> 16:15:21; Kweichow Moutai 600519: 251; Ping An Bank 000001: 248)	00700's latest snapshot at 16:15:21, about 70 seconds behind
`stock_kline`	Minute-bar OHLCV + turnover + tick_count	5 granularities (1m/5m/15m/30m/60m) x HK 2,866 / SZ 2,894 / SH 2,316 stocks	Minute-level	HK 1m bar latest at 16:07; SH/SZ latest at 15:00 (already closed for the day)
`fund_realtime (exchange ETF)`	Live price + IOPV (indicative NAV) + premium/discount rate + bid1/ask1 + turnover rate + market cap/float cap	1,556 exchange-traded ETFs; IOPV populated in ~96% of rows (2,918,471 / 3,034,976); premium/discount ~95%	Near real-time intraday	Latest snapshot at 16:12, about 4 minutes behind
`fund_realtime (LOF)`	Same field structure as above, but IOPV and premium/discount are entirely empty	390 LOFs	Near real-time intraday	Latest snapshot at 16:16, about 30 seconds behind
`stock_day`	Daily OHLC + dealtime (refresh timestamp)	21,823 stocks, 1,156 trading days (since 2021-11-15)	6-10 refreshes per day, not minute-level; A-share data mostly lands after the 15:00 close	Today's latest: HK 16:01, SH 15:35, SZ 15:36
`metal_day`	SGE precious-metal daily bars, including settlement price and open interest	10 contracts, ~24 years since 2002-10-31, 35,010 rows	Daily	Same-day (2026-07-29)
`fund_day`	Fund daily NAV	27,446 funds, 4,420,707 rows	Daily (T+1 disclosure)	Web (PG) covers about the past year, since 2025-07-15
`fund_dividend`	Fund dividend records	11,413 records across 4,309 funds, ex-dividend dates through 2026-07-31	Event-driven (only updates on a dividend event)	—
`fund_codes`	Fund static attributes: type / manager / custodian / fees / is_etf / is_lof	31,694 funds, of which 1,768 are exchange-traded ETFs	Daily sync	—

What counts as minute-level, and what doesn't — we say it upfront

Most competitors are vague about what "minute-level" or "real-time" actually means. We'd rather draw the line clearly — here are the three distinctions most worth checking before you rely on this data.

stock_minutes is a five-level snapshot, not tick-by-tick

It samples the last price and five-level bid/ask price & size roughly every 60-90 seconds — it is not exchange tick-by-tick trade data. If you need raw order flow, this isn't it; if you need minute-level candles, use stock_kline.

stock_kline is the real minute-level candlestick

OHLCV bars across five granularities (1m/5m/15m/30m/60m) — genuinely minute-level and usable directly for minute-level technical analysis or backtesting.

stock_day is not minute-level, even though it refreshes intraday

Measurements show A-share stock_day has only 6-10 distinct dealtime values per day, with the earliest refresh for SH/SZ at 15:00 (the closing auction) — meaning it refreshes 6-10 times a day and mostly lands after close. Don't treat it as an intraday minute-level signal.

Why exchange ETF traders should watch IOPV and premium/discount

IOPV (indicative optimized portfolio value) and the premium/discount rate are the two numbers exchange-traded ETF traders care about most — and the two most often missing from free data sources. They tell you whether the current trade price is running above or below the ETF's actual underlying NAV.

1,556 exchange ETFs, IOPV populated in ~96% of rows

We provide IOPV for 1,556 exchange-traded ETFs, populated in about 96% of rows measured (2,918,471 / 3,034,976); premium/discount rate is populated in about 95%.

LOFs currently have no IOPV or premium/discount

The 390 LOFs share the same real-time field structure as ETFs, but the IOPV and premium/discount columns are currently empty across the board — we'd rather state that gap plainly than borrow ETF field names to paper over it.

Why most free sources skip these two numbers

IOPV requires real-time valuation of an ETF's underlying constituents, which costs more to compute than simply forwarding the order-book price — which is why it's commonly omitted from free market apps.

Web-side retention windows — different per asset, not just "full history"

The web app and dashboards query a PG mirror directly, and the retention window differs by asset — from as short as 3 days to about 24 years. The table below lists each one; don't treat it as a blanket "we have everything."

Asset	PG-side measured retention window
`stock_day`	1,156 trading days (~4.6 years, since 2021-11-15)
`stock_kline`	9 days (20260721 -> 20260729)
`stock_minutes`	3 days (20260727 -> 20260729)
`fund_realtime`	8 days (20260722 -> 20260729)
`fund_day`	about 1 year (since 2025-07-15)
`metal_day`	full history, about 24 years

These are the PG-mirror windows queried directly by the web app / dashboards. Minute-level and intraday data (stock_kline / stock_minutes / fund_realtime) refresh more often and accumulate faster, so their web-side windows are correspondingly shorter. If you need longer history for backtesting or research, AI agents can query longer history hosted on MaxCompute (the MaxCompute start date varies by asset and can be queried on the agent side as needed).

Common questions about real-time / minute-level data coverage

Is stock_day updated at minute-level?

No. Measurements show A-share stock_day has only 6-10 distinct dealtime values per day, and the earliest refresh for SH/SZ is 15:00 (after the closing auction) — it refreshes 6-10 times a day and mostly lands after market close. Don't use it as an intraday minute-level signal. The genuinely minute-level assets are stock_kline and stock_minutes.

Is stock_minutes the same as tick-by-tick data?

No. stock_minutes is a five-level bid/ask snapshot roughly every 60-90 seconds (last price plus five-level bid/ask price and size), not exchange tick-by-tick trade data. If you need tick data, this asset isn't a fit; if you need minute-level candles, use stock_kline instead.

What's the IOPV coverage for exchange-traded ETFs, and why does it matter?

We provide IOPV for 1,556 exchange-traded ETFs, populated in about 96% of rows measured; premium/discount rate in about 95%. IOPV tells you whether the current trade price is running above or below the ETF's actual underlying NAV — the key number for deciding whether to trade at the current price, and something most free data sources don't provide.

Are these freshness numbers (like "70 seconds behind") a real-time guarantee?

No. These numbers come from a single sampling run at 2026-07-29 16:16, meant to illustrate the order of magnitude — not a real-time promise for every visit. Actual freshness varies with trading session, network conditions, and how actively an asset is traded; check the real timestamp you see after signing in.

The web-side retention window is short — how do I query longer history?

The web app and dashboards read a PG mirror, and the retention window differs by asset (stock_minutes 3 days, stock_kline 9 days, fund_realtime 8 days, stock_day about 4.6 years, metal_day full history of about 24 years). If you need longer history for backtesting or research, AI agents can query more complete history hosted on MaxCompute.

Is minute-level coverage the same for HK stocks and A-shares?

Close but not identical: stock_minutes covers A-share SH 2,463 / SZ 3,081 and HK 2,866; stock_kline's five minute granularities cover HK 2,866 / SZ 2,894 / SH 2,316. The exact coverage is whatever asset list you see when you subscribe.

Want to see what this data actually looks like?

Start by previewing sample data without registering, then sign in to connect your real holdings to AI portfolio analysis and risk monitoring.