Coverage
What you get
| Item | Detail |
|---|---|
| Product code | STK-TRADES |
| Coverage | US-listed stocks and ETFs, lit exchanges and off-exchange (TRF) prints |
| Granularity | One row per trade, nanosecond SIP timestamps |
| History | 10 years on Pro, full history on Business (planned) |
| Delivery | REST by symbol and day, WebSocket for live trades (Business), CSV and Parquet files |
| Update speed | 15-minute delayed on Pro, real-time SIP on Business |
Bars hide things. A 1-minute candle tells you the high was 228.60, but not whether that was one odd-lot print on a dark venue or 40,000 shares on Nasdaq. Tick data is the raw tape. Everything else (bars, VWAP, volume profiles) is built from it, which is why we keep it in the same schema as the rest of our stock products.
Schema
Schema
One row per trade. timestamp is the SIP time and is what you sort on. participant_timestamp is the venue's own clock, useful when you care about latency between the exchange and the consolidated feed.
| Field | Type | Required | Description |
|---|---|---|---|
symbol |
string | yes | Our ticker |
timestamp |
string | yes | SIP timestamp, ISO 8601 UTC, nanosecond precision |
participant_timestamp |
string | optional | Exchange's own timestamp |
price |
number | yes | Trade price |
size |
number | yes | Shares, including odd lots (fractional allowed) |
exchange |
string | yes | Venue MIC or our exchange code |
conditions |
array of strings | yes | Sale condition codes (e.g. @, F, T, I) |
tape |
string | yes | A (NYSE), B (Arca/other), C (Nasdaq) |
trade_id |
string | optional | Venue trade ID |
sequence_number |
number | yes | SIP sequence number for ordering |
is_off_exchange |
boolean | yes | True for TRF / dark pool prints |
trf_id |
string | optional | Trade reporting facility ID for off-exchange prints |
correction |
number | optional | Correction indicator |
bid_price |
number | optional | NBBO bid at trade time |
ask_price |
number | optional | NBBO ask at trade time |
aggressor_side |
string | optional | buy, sell, unknown |
session |
string | optional | pre, regular, post, overnight |
Example record:
{
"symbol": "AAPL",
"timestamp": "2026-11-18T14:30:01.204518337Z",
"participant_timestamp": "2026-11-18T14:30:01.204390112Z",
"price": 228.43,
"size": 100,
"exchange": "XNAS",
"conditions": ["@"],
"tape": "C",
"trade_id": "52983525029461",
"sequence_number": 1048231,
"is_off_exchange": false,
"bid_price": 228.42,
"ask_price": 228.44,
"aggressor_side": "buy",
"session": "regular"
}
Sample data
Sample data
There is no free stock tick sample yet. US trade data comes from the CTA and UTP plans, and we won't hand out a file before the licence that covers it is signed. You can still ask: tell us the symbol and the day you want, and we'll send a sample in the schema above once we can.
Use cases
Use cases
Custom bars. Time bars are one choice among many. With price, size and timestamp you can build volume bars, dollar bars or 10-second bars that no vendor ships, and resample them however your model wants.
Filtering out junk prints. Most bad backtests on tick data come from trades that should never have been in the sample: late reports, out-of-sequence prints, average-price trades. The conditions array and the correction field let you drop them before they set a fake high of day.
Dark pool share of volume. is_off_exchange and trf_id flag prints reported through a trade reporting facility. Sum size by that flag per day and you get the off-exchange share for any ticker, which many traders watch around earnings.
Order flow and aggressor side. When bid_price, ask_price and aggressor_side are present, you can tell whether a trade lifted the offer or hit the bid. Cumulative delta from that is a standard input for intraday momentum and absorption studies.
Execution checks. You got filled at 228.43 at 14:30:01. The tape shows what else printed in that second and on which venue (exchange, tape), so you can see whether the fill was fair.
Code
Code examples
Example: all trades for one symbol on one day.
curl "https://api.marketdata.co/v1/stocks/trades?symbol=AAPL&date=2026-11-18&limit=50000" \
-H "Authorization: Bearer $MARKETDATA_API_KEY"
Example: the same request in Python, then 5-second bars built from the trades.
import os
import requests
import pandas as pd
resp = requests.get(
"https://api.marketdata.co/v1/stocks/trades",
params={"symbol": "AAPL", "date": "2026-11-18", "limit": 50000},
headers={"Authorization": f"Bearer {os.environ['MARKETDATA_API_KEY']}"},
)
resp.raise_for_status()
ticks = pd.DataFrame(resp.json()["results"])
ticks["timestamp"] = pd.to_datetime(ticks["timestamp"])
ticks = ticks.set_index("timestamp").sort_index()
if "correction" in ticks: # drop corrected prints
ticks = ticks[ticks["correction"].fillna(0) == 0]
bars = ticks["price"].resample("5s").ohlc()
bars["volume"] = ticks["size"].resample("5s").sum()
print(bars.dropna().head())
Large days return more than one page. The response includes a next_url you follow until it is empty; see pagination.
Pricing
Pricing
| Plan | Stock tick data |
|---|---|
| Pro, $79/mo | 15-minute delayed trades by REST, 10 years of history |
| Business, $249/mo | Real-time trades over WebSocket, full history, commercial internal use, exchange fees passed through at cost |
Prefer files you keep? The planned dataset price is $15 per symbol-year, in CSV or Parquet. One full US market day is planned at $99, with a $49 minimum order. These are early access prices and may change before launch.
For context, the providers we studied charge $9.99 to $2,499 a month for plans that include stock trades, with a median of about $99 across 18 published prices. Four offer a limited free version.
Side by side
How we compare
22 of the 50 providers we studied sell US stock trades, so this is not a rare product. What's rare is buying a slice of it. Most sell tick history only as part of a monthly plan, and the file vendors that do sell one-off orders tend to have high minimums. We plan to sell a single symbol-year as a file you keep, publish the price on this page, and use the same field names (symbol, timestamp, price, size) as our crypto and futures trade files, so one loader reads all of them.
FAQ
Frequently asked questions
How far back does stock tick history go?
Pro is planned with 10 years of trade history and Business with the full history of our source. Exact start dates per symbol will be published before launch, once the feed is signed.
Are tick prices adjusted for splits?
No. Trades are stored exactly as they printed, because adjusting a tape would change real fills. If you need split-adjusted prices, apply the factors from our corporate actions data or use our bar products.
What time zone are the timestamps in?
UTC, in ISO 8601 with nanosecond precision, for example 2026-11-18T14:30:01.204518337Z. That is 09:30 New York time on that date. Our timestamps guide covers the daylight saving shift.
Is the data real-time or delayed?
Pro is planned as 15-minute delayed and Business as real-time SIP. Real-time US trade data carries exchange fees, which is why it sits on the higher plan.
Which formats can I download?
CSV and Parquet for dataset files, JSON from the API. Parquet is the better choice for tick data, since a busy symbol can run to millions of rows a day.
Do you include pre-market and after-hours trades?
Yes. The session field marks each trade as pre, regular, post or overnight, so you can filter to regular hours in one line.
Keep reading
Related
- Asset hub: Stocks
- Symbols: AAPL, MSFT, NVDA, AMZN, GOOGL, META
- Docs: API reference: stock trades and quotes
- Learn: Tick data vs bar data, Dark pools explained