IRONFLOW
SANDBOXDATA APILIVEPRICINGDOCSGET_API_KEY
GET_API_KEY
GUIDES · NOVEMBER 2026 · 3 MIN READ

Freqtrade on Hyperliquid: backtest past 5,000 candles

Freqtrade's own Hyperliquid notes say it: Hyperliquid only offers 5,000 historic candles, so a backtest is limited to the last 5,000. Write longer history into Freqtrade's data folder and backtest as usual: free hourly candles and funding since May 2025, or 1m to 30m bars with a free key.

FILE
huggingface.co/datasets/ironflow/hyperliquid-candles-1h free, hourly
API
GET https://api.ironflow.sh/v1/candles?market=BTC-PERP&interval=5m
HOW
docs.ironflow.sh/guides/freqtrade-hyperliquid-historical-data

Free key14 days of history at 60 requests a minute. Without a key, the API reaches 24 hours back.

By requesting a key you agree to the Terms and API terms; see the Privacy Policy. We may email you when your key keeps hitting its limits; you can opt out at any time.

CONTENTS
  1. The short version
  2. Why Freqtrade stops at 5,000
  3. What to load instead
  4. Check the data
  5. Load it into Freqtrade
  6. Questions
  7. Related guides
  8. References

The short version#

  • The cap: freqtrade download-data gets at most the newest 5,000 candles from Hyperliquid. At 1h that is about 208 days.
  • The fix: write candles and funding from another source as Freqtrade feather files, then run freqtrade backtesting as usual.
  • How much: 12,586 free hourly BTC candles since 25 May 2025, 2.5 times the exchange's 5,000.

Why Freqtrade stops at 5,000#

Hyperliquid's candleSnapshot keeps only the most recent 5,000 candles per interval (info endpoint). Freqtrade sets ohlcv_has_history to False for Hyperliquid, and its exchange notes say backtests will either need to build candles over time or be limited to the last 5,000.

Timeframe5,000 candles reach back
1mAbout 3.5 days
5mAbout 17 days
15mAbout 52 days
1hAbout 208 days
4hAbout 833 days

What to load instead#

SourceCoversSizeAccess
Free hourly candles25 May 2025 to 31 Oct 202612,586 BTC candlesNo key
Free fundingMay 2023 to Oct 2026Hourly ratesNo key
/v1/candles, free keyLast 14 days20,160 1m candles60 requests a minute
/v1/candles, ExplorerLast 90 days1m to 1w$49 for 30 days
/v1/candles, BuilderLast year1m to 1w$149 for 30 days

Check the data#

Count what the free hourly file holds, no key:

python
import pandas as pd

URL = "https://huggingface.co/datasets/ironflow/hyperliquid-candles-1h/resolve/main/data/{}.parquet"
months = pd.period_range("2025-05", "2026-10", freq="M").strftime("%Y-%m")
candles = pd.concat(pd.read_parquet(URL.format(m)) for m in months)

btc = candles[candles.coin == "BTC"]
print(len(btc), "hourly BTC candles from", btc.timestamp.min(), "to", btc.timestamp.max())
text
12586 hourly BTC candles from 2025-05-25 14:00:00+00:00 to 2026-10-31 23:00:00+00:00

Load it into Freqtrade#

The Freqtrade guide in our docs has two scripts: free hourly candles and funding with no key, and 1m to 30m bars from /v1/candleswith a key. Both write with Freqtrade's own data handler to user_data/data/hyperliquid/futures, pairs named BTC/USDC:USDC. After the hourly script, with Freqtrade 2026.10:

bash
freqtrade list-data --datadir user_data/data/hyperliquid --trading-mode futures --show-timerange
text
┃          Pair ┃ Timeframe ┃         Type ┃                From ┃                  To ┃ Candles ┃
│ BTC/USDC:USDC │        1h │ funding_rate │ 2025-06-01 00:00:00 │ 2026-10-31 23:00:00 │   12432 │
│ BTC/USDC:USDC │        1h │      futures │ 2025-06-01 00:00:00 │ 2026-10-31 23:00:00 │   12432 │
│ ETH/USDC:USDC │        1h │ funding_rate │ 2025-06-01 00:00:00 │ 2026-10-31 23:00:00 │   12432 │
│ ETH/USDC:USDC │        1h │      futures │ 2025-06-01 00:00:00 │ 2026-10-31 23:00:00 │   12432 │
  • In your config: "trading_mode": "futures", a margin_mode (Freqtrade supports isolated and cross on Hyperliquid) and "stake_currency": "USDC".
  • Then run freqtrade backtesting with a --timerange inside those dates.
  • To test the same idea without Freqtrade, see how to backtest on Hyperliquid.

Questions#

Why can't Freqtrade download Hyperliquid history?#

Hyperliquid's candleSnapshot keeps only the most recent 5,000 candles per interval, and Freqtrade marks Hyperliquid as having no OHLCV history. Freqtrade's docs say a backtest is limited to the last 5,000 candles unless you build the candles yourself over time.

How do I backtest Freqtrade on Hyperliquid for more than 5,000 candles?#

Write longer history into Freqtrade's data folder as futures and funding_rate feather files, then run freqtrade backtesting as usual. Ironflow's free hourly candles reach back to 25 May 2025: 12,586 BTC candles, against 5,000 from the exchange.

Can I get 1m or 5m Hyperliquid candles for Freqtrade?#

Yes, from GET https://api.ironflow.sh/v1/candles with a key. A free key covers 14 days, which is 20,160 1m candles. Explorer covers 90 days and Builder one year.

Does the backtest include Hyperliquid funding?#

Yes, if you write the funding_rate file too. The free funding dataset has hourly rates from May 2023, and the docs guide stores them next to the candles.

Related guides#

  • Hyperliquid candles and OHLCV data: the 5000 candle limit and older history
  • How to backtest on Hyperliquid
  • How to build a Hyperliquid trading bot in Python

References#

  • Freqtrade docs: Hyperliquid. The 5,000 candle note.
  • Hyperliquid docs: Info endpoint. candleSnapshot and its 5,000 candle limit.
  • Hugging Face: ironflow/hyperliquid-candles-1h. The free hourly candles.
  • Hugging Face: ironflow/hyperliquid-funding. The free hourly funding rates.

For 1m to 30m bars, get a free key: 14 days, by email, no card.

Free key14 days of history at 60 requests a minute. Without a key, the API reaches 24 hours back.

By requesting a key you agree to the Terms and API terms; see the Privacy Policy. We may email you when your key keeps hitting its limits; you can opt out at any time.

ON THIS PAGE
  1. The short version
  2. Why Freqtrade stops at 5,000
  3. What to load instead
  4. Check the data
  5. Load it into Freqtrade
  6. Questions
  7. Related guides
  8. References

Automated derivatives trading is high-risk: you can lose all of your funds. Ironflow is trading software, not financial advice, and never holds your funds. Not available to U.S. persons or in sanctioned jurisdictions. See Risk Disclosure.

Not affiliated with Hyperliquid or Pear Protocol.

IRONFLOW

© 2026 IRONFLOW

Product

IronboxAPI and MCPPricingNodesTerminal

Data

Live liquidationsLiquidation mapLiquidation heatmapHIP-4 outcomesTop tradersDatasets

Developers

DocsGuidesFree API keyStatusGitHub

Company

TermsPrivacyRiskAcknowledgements