Get Free Forex Data in Python — No API Key Required

Most "free" forex APIs cap you at a few hundred requests a day and make you register for a key first. biquote is different: every endpoint is anonymous, and the limit is 15,000 requests per minute. Here is everything you need.

1. A live quote in one line

import requests

tick = requests.get("https://biquote.io/api/EURUSD").json()
print(tick["mid"], tick["dayDiffPercent"])   # 1.15884 0.51

Or with the client:

from biquote import Biquote

bq = Biquote()
print(bq.tick("EURUSD")["mid"])

2. Many symbols in one request

Polling one symbol per call wastes your quota. The batch endpoint takes up to 100 symbols (repeat the parameter — the comma form is not supported):

r = requests.get(
    "https://biquote.io/api/latest",
    params=[("symbols", s) for s in ("EURUSD", "XAUUSD", "BTCUSD", "USDTRY")],
).json()
for sym, t in r.items():
    print(f'{sym:8} {t["mid"]:>12}  {t["dayDiffPercent"]:+.2f}%')

3. OHLC candles into pandas

import pandas as pd

env = requests.get(
    "https://biquote.io/api/XAUUSD/ohlc",
    params={"interval": "1h", "limit": 500},
).json()
df = pd.DataFrame(env["bars"])            # envelope: {symbol, interval, bars}
df["openTime"] = pd.to_datetime(df["openTime"])
df = df.set_index("openTime").sort_index()  # bars arrive newest-first
print(df[["open", "high", "low", "close"]].tail())

Intervals: 1m 5m 15m 30m 1h 4h 1d, up to 1,000 bars per call.

4. Streaming instead of polling

For continuous updates, skip HTTP entirely — the SignalR hub pushes every tick with no rate limit:

pip install pysignalr

import asyncio
from pysignalr.client import SignalRClient

async def main():
    client = SignalRClient("https://biquote.io/hubs/tick")
    client.on("ReceiveTick", lambda t: print(t[0]["symbol"], t[0]["mid"]))
    client.on_open(lambda: client.send("Subscribe", [["EURUSD", "XAUUSD"]]))
    await client.run()

asyncio.run(main())

Which symbols are live?

The catalogue lists ~1,700 instruments but only ~280 quote at any moment. Always filter:

live = requests.get("https://biquote.io/api/symbols",
                    params={"quotedWithinDays": 7}).json()
print(len(live), "instruments")

quotedWithinDays rather than liveOnly: the latter means "quoting this second", so at a weekend it returns almost nothing and any list built on it comes back empty.

Rate limits & errors

That last row is the one that catches people. Forex stops quoting from Friday night to Sunday night, and a closed market is not an error — you get the last known price plus marketState, stale and quoteAgeSeconds. Show it as a last price rather than a live one, and hide the day change, which belongs to a session that has already ended:

tick = requests.get("https://biquote.io/api/EURUSD").json()

if tick["marketState"] == "closed":
    print(f"{tick['mid']}  (market closed, {tick['quoteAgeSeconds'] // 3600}h ago)")
else:
    print(f"{tick['mid']}  {tick['dayDiffPercent']:+.2f}%")

Writing a trading client instead? Pass allowStale=false and you get a 404 the moment a quote is more than five minutes old — the right behaviour when acting on a stale price is worse than acting on none.

Read the full API reference →