What Is Freqtrade? How the Open-Source Crypto Trading Bot Works

Freqtrade is a free, open-source trading bot for crypto exchanges, written in Python. You give it a strategy (a Python file with rules for when to buy and sell), an exchange account with an API key, and a configuration file. It then watches the market and places orders for you, around the clock.

It is one of the most widely used open-source projects in crypto trading, with tens of thousands of stars on GitHub and an active Discord community. The software is free under the GPL-3.0 licence. There is no subscription and no company taking a cut of your trades.

What freqtrade actually does

Every few seconds, freqtrade:

Downloads the latest candles (open, high, low, close, volume) for the coins in your list.
Runs your strategy, which calculates indicators such as RSI or moving averages and marks candles where it wants to enter or exit.
Places, tracks and cancels orders on the exchange according to those signals and your risk settings.
Records every trade in a local database, so you can see profit, open positions and history.

You control it through a Telegram bot, a built-in web interface called FreqUI, or the command line.

The parts of a strategy

A freqtrade strategy is a Python class with a few required pieces:

Part	What it does
`timeframe`	Candle size the strategy reads, such as `5m` or `1h`
`populate_indicators()`	Adds indicators to the price data
`populate_entry_trend()`	Marks candles where the bot should buy
`populate_exit_trend()`	Marks candles where the bot should sell
`minimal_roi`	Take-profit levels, which can fall over time
`stoploss`	Maximum loss per trade before the bot exits

Here is the entry rule from BbandRsi, one of the strategies tested on this site:

dataframe.loc[
    (dataframe['rsi'] < 30) &
    (dataframe['close'] < dataframe['bb_lowerband']),
    'enter_long'] = 1

In words: buy when RSI is below 30 and price has closed below the lower Bollinger Band.

Tools for testing before you trade

Freqtrade's biggest strength is its testing toolkit. You should use all of it before risking money.

Backtesting replays your strategy over historical data and reports profit, drawdown, win rate and more. See our backtesting tutorial.
Hyperopt searches for strategy parameters that performed best in the past. It is powerful and easy to misuse; our hyperopt guide covers how to avoid over-fitting.
Dry-run trades live market data with simulated money. It is the closest thing to real trading without the risk.
FreqAI adds machine-learning models that retrain themselves as new data arrives. It is an advanced module and needs more computing power.

What you need to run it

A computer that stays on. Most people use a small Linux VPS; see choosing a VPS for freqtrade. The freqtrade README recommends at least 2 GB of RAM, 2 CPU cores and 1 GB of disk.
Docker (recommended) or Python 3.12 or newer. Our Docker installation guide takes about 15 minutes.
An account on a supported exchange. Freqtrade officially supports Binance, Bybit, Bitget, OKX, Gate, HTX, Kraken, BingX and Hyperliquid for spot, and a subset of these for futures.
An API key with trading permission. Never enable withdrawals on a bot's API key.
An accurate system clock. Exchanges reject requests from machines whose time drifts, so keep NTP time sync on.

What it costs

The software is free. Your real costs are:

A VPS, typically a few dollars a month for a small instance.
Exchange trading fees on every buy and sell. With a 0.1% fee per side, a strategy that trades several times a day pays a lot in fees, which is why every backtest on this site includes them.
Your time. Freqtrade rewards people who are comfortable editing configuration files and reading Python.

Does freqtrade make money?

Freqtrade is a tool. Whether it makes money depends entirely on the strategy and the market. Our leaderboard tests 27 community strategies on identical data and settings. Over our two-year window, which included a strong bull market, none of them beat simply holding Bitcoin, and more than half lost money.

That does not mean automated trading cannot work. It means the free sample strategies are learning material, not finished money-makers, and that any strategy you run deserves serious testing.

The freqtrade project's own README describes the software as being for educational purposes and tells users not to risk money they are afraid to lose. We agree.

Freqtrade compared with paid bot platforms

Hosted platforms such as 3Commas, Cryptohopper or Pionex give you a web interface and preset bots for a monthly fee or built-in costs. Freqtrade gives you full control and transparency, no subscription and your keys stay on your own server. In exchange, you handle the setup, updates and debugging yourself. Our full comparison covers the differences in detail.

Where to go next

Install freqtrade with Docker
Connect it to Bybit
Backtest a strategy
Browse tested strategies and pick one to study

Ready to connect an exchange?

Freqtrade needs an exchange account with API access. Bybit supports both spot and futures in freqtrade and offers a demo mode for testing.

Open a Bybit account Browse tested strategies

Affiliate link: we may earn a commission if you sign up, at no cost to you.

What is Freqtrade? A plain-English guide to the open-source crypto bot