Crypto Tick Data: Tick-by-Tick Trade History for Every Venue and Instrument
Tick data here means every trade the venue printed — price, size, aggressor side, trade id, the exchange timestamp and our co-located receive timestamp in nanoseconds — not a sampled tape and not aggregated trades from a REST backfill. In the 30 days to 2026-09-16 the archive recorded 19.0B trades worth $5.7T across 32 venues, about $297 per trade.
Crypto Tick Data — Tick-by-Tick Trade History
A trade in a day file is a type-2 message: the venue's own trade id, price, quantity, whether the aggressor was a buyer or seller, and two timestamps. Because the recorder subscribes to the full public feed rather than polling a trades endpoint, bursts are complete — thousands of prints in one millisecond arrive in order, with the event-id chain proving nothing was dropped.
The tape sits next to the order book in the same file, so every trade can be placed against the book state that produced it: which level it hit, how much depth was left, whether it was a liquidation (derivatives files carry those as their own type-17 events, priced at the bankruptcy price). That pairing is what market-impact, toxicity and fill-probability research needs and what a trades-only feed cannot give.
Trades export flat as gzipped CSV or Parquet (?format=trades.csv.gz, ?format=trades.parquet) on every sample and purchased day — columns per the docs convention, timestamps in integer microseconds UTC, ready for pandas, polars or DuckDB. The native file keeps the nanoseconds.
Venues ranked by recorded trade count — last 30 days
19.0B trade prints across 31 venues in the 30 days to 2026-09-16 — about 634.1M per UTC day, $297 per trade on average. Counted from the files, not from venue-reported volume.
| # | Venue | Trades · 30d | Trades / day | Avg trade | Share of prints |
|---|---|---|---|---|---|
| 1 | BingX | 6.7B | 222.6M | $26 | 35.1 % |
| 2 | Binance USDT-M | 5.2B | 173.9M | $339 | 27.4 % |
| 3 | Bybit | 1.6B | 52.9M | $314 | 8.3 % |
| 4 | Binance Spot | 1.1B | 36.9M | $276 | 5.8 % |
| 5 | Okx | 865.1M | 28.8M | $941 | 4.5 % |
| 6 | Bitget | 530.6M | 17.7M | $547 | 2.8 % |
| 7 | Gate.io Future | 422.0M | 14.1M | $874 | 2.2 % |
| 8 | Mexc | 359.3M | 12.0M | $143 | 1.9 % |
| 9 | Whitebit | 323.5M | 10.8M | $780 | 1.7 % |
| 10 | Kalshi | 251.2M | 8.4M | $41 | 1.3 % |
| 11 | HTX | 229.7M | 7.7M | $103 | 1.2 % |
| 12 | Gate.io Spot | 201.9M | 6.7M | $268 | 1.1 % |
| 13 | Bitmart | 195.7M | 6.5M | $858 | 1.0 % |
| 14 | KuCoin Futures | 193.7M | 6.5M | $156 | 1.0 % |
| 15 | Coinbase International | 161.0M | 5.4M | $1,722 | 0.8 % |
Trade prints per venue over the 30 days to 2026-09-16, counted from the tick files (every print, all instrument classes); trades / day divides by the venue's recorded days, avg trade = USD turnover ÷ prints.
Recorded trades and turnover per month — all venues
| Month | Turnover (USD) | Trades | Avg / day | Days | vs previous month |
|---|---|---|---|---|---|
| Sep 2026 (month to date) | $2.8T | 9.8B | $172.7B | 16 | — |
| Aug 2026 | $4.7T | 18.7B | $153.1B | 31 | +11.4 % |
| Jul 2026 | $4.3T | 19.3B | $137.4B | 31 | −20.4 % |
| Jun 2026 | $5.4T | 22.0B | $178.4B | 30 | +11.6 % |
| May 2026 | $4.8T | 20.2B | $154.6B | 31 | +0.1 % |
| Apr 2026 | $4.8T | 20.5B | $159.7B | 30 | −14.6 % |
| Mar 2026 | $5.6T | 20.6B | $181.0B | 31 | −5.5 % |
| Feb 2026 | $5.9T | 22.0B | $212.0B | 28 | −26.5 % |
| Jan 2026 | $8.1T | 25.1B | $260.6B | 31 | +32.3 % |
| Dec 2025 | $6.1T | 26.2B | $197.1B | 31 | −25.0 % |
| Nov 2025 | $8.1T | 30.2B | $271.5B | 30 | −22.6 % |
| Oct 2025 | $10.5T | 31.3B | $339.5B | 31 | — |
All venues — USD turnover and trade counts summed per calendar month (UTC) from our tick archive. Days = UTC days with recorded activity; a partial month shows no month-over-month change.
Message types behind trades & ticks data
| Type | Message | What it carries |
|---|---|---|
2 | Trades | Every print with aggressor side, price, size, trade id and exchange timestamp |
17 | Liquidations | Derivatives, 2026 files onward, priced at the bankruptcy price |
How to download trades & ticks data
- Find the instrument in the Data Shop (search by pair or venue) or browse by exchange, by asset or by trading pair.
- Try a free day first: the full-day samples on the downloads page are the exact shop files, no account needed — parse one before you buy a month.
- Pick the UTC days in the instrument's calendar and check out — card with invoice, or Premium and pack credits (1 credit = €1 = one instrument-day). Many instruments at once: "Select to buy" in the instrument scanner adds a shared date window.
- Export: append
?format=trades.csv.gzor?format=trades.parquetto any download link for a flat trades table,?format=liquidations.csv.gzon derivatives days for the liquidation prints, or read the native zstd JSON-lines file with the free Python reader. - Download from the order page (single files), your account (batch ZIP of everything you own) or scripted via the MCP server and the Agent Skill. Days in deep archive restore in about 12 hours.
Research, backtesting and model training
Purchased data is for your own use — backtesting, research, analytics and training your own models, in-house or in your own products; the trained models and derived data you build from it are yours. The raw files and tick streams themselves may not be redistributed, resold or published. The full terms are on the data license page.
Trades & ticks data — FAQ
What is in a crypto tick data file?
Every trade print of one instrument for one UTC day — price, size, aggressor side, the venue's trade id, exchange timestamp and receive timestamp in nanoseconds — plus, in the same file, the full Level-2 order-book history, top-of-book quotes where the venue publishes them, and on derivatives funding, mark/index prices and liquidation events.
Is the trade history complete or sampled?
Complete. The recorder is co-located at the venue and writes the public trade stream as it arrives; bursts of thousands of prints per millisecond are kept in order, and an event-id chain makes any gap visible. The free reader's audit command reports chain breaks before you build on a file.
What timestamp resolution does the tick data have?
Integer nanoseconds for both the exchange timestamp and our co-located receive timestamp in the native file; the flat CSV/Parquet exports carry integer microseconds UTC. Whether a venue's own timestamp is millisecond- or microsecond-granular depends on its feed — check the current docs of the venue; the receive timestamp is ours and always nanosecond.
Does the tick data include the trade side (buyer or seller aggressor)?
Yes — every trade carries the aggressor side as the venue reports it, and the 1-minute statistics derived from the files split trade counts and USD turnover into buy and sell. Liquidation trades on derivatives are additionally flagged and, from 2026 files onward, recorded as their own events at the bankruptcy price.
How is this different from downloading trades from the exchange API?
Exchange REST endpoints return recent trades in pages with limits that vary per venue (check the current docs), often without the aggressor side or with millisecond timestamps, and never the book state next to each trade. The archive is the live feed recorded end to end at the venue, normalized into one schema across every exchange, sold per instrument-day.
How many trades does the archive record per day?
About 634.1M trade prints per UTC day across 31 venues in the 30 days to 2026-09-16, averaging $297 per trade. Counted from the tick files themselves — every print, not venue-reported volume.
Which crypto exchange prints the most trades?
1. BingX (6.7B trades, 222.6M/day), 2. Binance USDT-M (5.2B trades, 173.9M/day), 3. Bybit (1.6B trades, 52.9M/day) in the 30 days to 2026-09-16 — 71 % of all prints in the archive. A high count with a small average trade size means a retail-heavy tape; the per-venue table on this page shows both.