Python OHLC API Guide

Quick start with requests

Install requests and pandas, then call the OHLC endpoint with a symbol, timeframe and date range. The data key contains the bar array. Always call raise_for_status() so a failed request does not silently become an empty DataFrame.

Chunk large date ranges

One request can return at most 50,000 rows and cover at most 366 days. For a multi-year M1 or M5 history, loop over monthly or quarterly windows. Keep the chunk boundaries explicit so you can resume after a failure.

• Use monthly chunks for M1 and M5.
• Use quarterly or yearly chunks for M15 and higher.
• De-duplicate on Timestamp after concatenation.
• Cache raw responses locally before transforming them.

Handle 429 and 413 correctly

A 429 response includes Retry-After. Wait that many seconds before retrying. A 413 means the requested row count is too large; narrow the date range or use a higher timeframe. Do not retry a 413 unchanged.

Check data quality in Python

Call the quality endpoint before training a model. A simple rule is to reject grades D/F or any report with a critical issue, then inspect the warning list for zero volume, spread or gap problems.

Python research workflows

Request examples

import requests
import pandas as pd

url = "https://www.algospecial.com/api/ohlc.php"
params = {
    "symbol": "XAUUSD",
    "tf": "M15",
    "from": "2026-01-01",
    "to": "2026-03-31",
}
r = requests.get(url, params=params, timeout=30)
r.raise_for_status()
payload = r.json()
df = pd.DataFrame(payload["data"])
df["Timestamp"] = pd.to_datetime(df["Timestamp"])
print(payload["rows"])
print(df.head())

import time
import requests
import pandas as pd
from datetime import date

url = "https://www.algospecial.com/api/ohlc.php"
frames = []

for year in range(2024, 2027):
    params = {
        "symbol": "EURUSD",
        "tf": "M15",
        "from": f"{year}-01-01",
        "to": f"{year}-12-31",
    }
    r = requests.get(url, params=params, timeout=30)
    if r.status_code == 429:
        wait = int(r.headers.get("Retry-After", "60"))
        time.sleep(wait)
        r = requests.get(url, params=params, timeout=30)
    r.raise_for_status()
    frames.append(pd.DataFrame(r.json()["data"]))
    time.sleep(0.5)

df = pd.concat(frames, ignore_index=True)
df["Timestamp"] = pd.to_datetime(df["Timestamp"])
df = df.drop_duplicates("Timestamp").sort_values("Timestamp")
print(len(df))

import requests

r = requests.get(
    "https://www.algospecial.com/api/data-quality.php",
    params={"symbol": "EURUSD", "tf": "M15"},
    timeout=30,
)
r.raise_for_status()
q = r.json()

if q["grade"] in ("D", "F") or q["critical_count"] > 0:
    raise SystemExit(f"Rejected: {q['grade']} {q['score']} {q['report_url']}")

print("Accepted", q["grade"], q["score"])

All timestamps are broker server time. Some legacy forex exports return 0 for TickVolume and Spread because those fields were not present in the source feed.

Related API guides