Bars

A bar is one OHLCV record: open, high, low, close, volume, VWAP and trade count over one interval. Each source stores its finest timeframe (/api/v1/timeframes?source=stocks lists it); any coarser timeframe is built from it when you ask, in UTC-aligned buckets.

Query

curl

curl -sD - "https://getmarkethistory.com/query/v1/bars/stocks/SPY/1d?start=2024-01-02&end=2024-01-31&adjust=all" \
  -H "Authorization: Bearer $MARKET_HISTORY_KEY"

Python client

from markethistory import Client

mh = Client()  # reads MARKET_HISTORY_KEY

df = mh.frame("stocks", "SPY", "1d", start="2024-01-02", end="2024-01-31", adjust="all")
df[["open", "high", "low", "close", "volume"]].head()

requests

import json
import os

import requests

url = "https://getmarkethistory.com/query/v1/bars/stocks/SPY/1d"
params = {"start": "2024-01-02", "end": "2024-01-31", "adjust": "all"}
headers = {"Authorization": f"Bearer {os.environ['MARKET_HISTORY_KEY']}"}

rows = []
while True:
    r = requests.get(url, params=params, headers=headers, timeout=60)
    r.raise_for_status()  # see /docs/errors for what each status means
    rows += [json.loads(line) for line in r.text.splitlines() if line]
    print(r.headers["X-Row-Count"], "rows,", r.headers["X-Bars-Charged"], "bars")
    if "X-Next-Cursor" not in r.headers:
        break
    params["cursor"] = r.headers["X-Next-Cursor"]

Response

HTTP/2 200
content-type: application/x-ndjson
x-row-count: 25
x-bars-charged: 100
x-query-ms: 38

{"close":458.21,"gap_fill_method":0,"high":462.37,"is_gap":false,"low":455.87,"open":461.53,"symbol":"SPY","symbol_id":1529,"timestamp":"2024-01-02T00:00:00Z","trade_count":701844,"volume":123623727.6,"vwap":458.996}
{"close":454.24,"gap_fill_method":0,"high":457.31,"is_gap":false,"low":453.78,"open":457.92,"symbol":"SPY","symbol_id":1529,"timestamp":"2024-01-03T00:00:00Z","trade_count":690121,"volume":103585820.9,"vwap":455.402}

Shape and headers as served; the values are illustrative, not copied from a response.

GET /query/v1/bars/{source}/{ticker}/{timeframe}, with a key.

Parameter	Meaning
source	The dataset (stocks, fx, crypto, …). /api/v1/sources lists them.
ticker	The symbol, as /api/v1/symbols lists it.
timeframe	1s, 1m, 5m, 1h, 1d, … The stored one or any coarser one.
start, end	Inclusive range (RFC 3339 or a bare date; a bare end is that whole UTC day).
adjust	raw (default), splits or all (splits and cash dividends). stocks only.
columns	Only these columns, comma-separated.
limit	Rows per page: default 1,000, maximum 100,000.
cursor	X-Next-Cursor from the previous page.
format	jsonl (default), json, csv, parquet, arrow or mbt1. Every one costs the same.

Pages run oldest first. Follow X-Next-Cursor (or the Link header) until it is absent. The client’s frame() and table() do it for you and fetch Parquet, so a long range arrives as typed columns (install it with the pandas and arrow extras).

adjust= rewrites open, high, low, close, vwap and volume from the published corporate actions; the stored bars stay raw. Keep the raw bars when you need to reproduce a backtest exactly: an adjusted history changes each time a new split or dividend lands.

What a 1d bar of US equities is

A UTC day of minute bars, pre- and after-hours included. Its close is the last trade before 00:00 UTC (20:00 New York in summer, 19:00 in winter), not the 16:00 closing print. In winter the last after-hours hour of a Friday falls on Saturday UTC, which makes a small Saturday bar. Drop it, or build daily bars from 1h in New York time; the moving-average guide does both.

Free sample

Hourly bars of any ticker over the last five years, with no key. With no dates you get the newest 100; with start, end or cursor it pages like the billed path, at most 5,000 bars a page. Other timeframes and adjust= are a 400 there.

curl

curl -sD - "https://getmarkethistory.com/api/v1/bars/stocks/AAPL/1h"

Python client

from markethistory import Client

page = Client().preview("stocks", "AAPL")  # no key
page.to_pandas().tail()

requests

import json

import requests

url = "https://getmarkethistory.com/api/v1/bars/stocks/AAPL/1h"
params = {}
headers = {}  # the free sample needs no key

rows = []
while True:
    r = requests.get(url, params=params, headers=headers, timeout=60)
    r.raise_for_status()  # see /docs/errors for what each status means
    rows += [json.loads(line) for line in r.text.splitlines() if line]
    if "X-Next-Cursor" not in r.headers:
        break
    params["cursor"] = r.headers["X-Next-Cursor"]

Response

HTTP/2 200
content-type: application/x-ndjson
x-row-count: 100
x-preview-limit: 100
x-preview-from: 2021-11-01T07:20:48Z
x-data-through: 2026-10-31T04:58:00Z

{"close":348.4375,"gap_fill_method":0,"high":348.47,"is_gap":false,"low":347.94,"open":347.98,"symbol":"AAPL","symbol_id":41,"timestamp":"2026-10-22T22:00:00Z","trade_count":1656,"volume":151877.0,"vwap":348.2574253059383}
{"close":348.6,"gap_fill_method":0,"high":348.75,"is_gap":false,"low":348.1428,"open":348.4,"symbol":"AAPL","symbol_id":41,"timestamp":"2026-10-22T23:00:00Z","trade_count":1679,"volume":40918.0,"vwap":348.491937076299}

Columns

timestamp (the bar’s open, UTC), open, high, low, close, volume, vwap, trade_count, symbol and symbol_id. is_gap and gap_fill_method are always false and 0: a missing interval is absent, never filled. Types and encodings are in Data & formats.

Price

One bar is one bar. A page rounds up to a block of 100, and an empty page is free. A free account’s monthly bars pay for 1h only; every other timeframe needs bought bars.