Loading...
Freqtrade-MCP bridges AI agents to the Freqtrade crypto bot via its REST API, enabling automated trading strategies and real-time data analysis.
Boost this tool
Subscribe to listing upgrades or segmented pushes.
Freqtrade-MCP bridges AI agents to the Freqtrade crypto bot via its REST API, enabling automated trading strategies and real-time data analysis.
This MCP server provides access to a live trading bot, so caution is warranted. It's relatively safe if Freqtrade's API is secured and credentials are well-managed. The risk is higher if the API is exposed without proper authentication or if the trading strategies are poorly designed, potentially leading to financial losses.
Performance is primarily limited by the Freqtrade API's response times and the network latency between the MCP server and the Freqtrade instance. Consider optimizing network configuration for low latency.
Cost considerations include the resources required to run the Freqtrade instance and the potential trading fees associated with placing trades. Also, consider the cost of API calls if Freqtrade implements rate limiting.
pip install freqtrade-clientfetch_market_dataRetrieves OHLCV (Open, High, Low, Close, Volume) data for a specified trading pair and timeframe.
Read-only operation; no modification of system state.
fetch_bot_statusGets the current status of the Freqtrade bot, including open trades and operational state.
Read-only operation; no modification of system state.
fetch_profitRetrieves a summary of the profit generated by the Freqtrade bot.
Read-only operation; no modification of system state.
fetch_balanceGets the current account balance from the Freqtrade bot.
Read-only operation; no modification of system state.
fetch_performanceRetrieves performance metrics for the trading bot.
Read-only operation; no modification of system state.
fetch_whitelistGets the list of trading pairs that the bot is allowed to trade.
Read-only operation; no modification of system state.
fetch_blacklistGets the list of trading pairs that the bot is blocked from trading.
Read-only operation; no modification of system state.
fetch_tradesRetrieves the history of trades executed by the bot.
Read-only operation; no modification of system state.
fetch_configGets the current configuration of the Freqtrade bot.
Read-only operation; no modification of system state.
fetch_locksRetrieves the list of active trade locks.
Read-only operation; no modification of system state.
place_tradePlaces a buy or sell trade for a specified trading pair.
Executes trades, potentially leading to financial gains or losses.
start_botStarts the Freqtrade trading bot.
Starts the bot, which will begin automated trading.
stop_botStops the Freqtrade trading bot.
Stops the bot, halting automated trading.
reload_configReloads the Freqtrade bot's configuration.
Reloads the bot's configuration, potentially changing its behavior.
add_blacklistAdds a trading pair to the blacklist, preventing the bot from trading it.
Modifies the bot's trading behavior by preventing trades on a specific pair.
delete_blacklistRemoves a trading pair from the blacklist, allowing the bot to trade it again.
Modifies the bot's trading behavior by allowing trades on a specific pair.
delete_lockDeletes a trade lock, potentially allowing a trade to execute.
Removes a trade lock, potentially affecting trading behavior.
Environment Variable
This MCP server provides access to a live trading bot, so caution is warranted. It's relatively safe if Freqtrade's API is secured and credentials are well-managed. The risk is higher if the API is exposed without proper authentication or if the trading strategies are poorly designed, potentially leading to financial losses.
The level of autonomy depends on how the tools are used. Read-only tools are safe for autonomous operation, while write tools should be used with caution and proper risk management strategies.
Production Tip
Implement robust error handling and monitoring within your AI agent to handle potential API errors and unexpected trading outcomes.
You need a running Freqtrade instance with the REST API enabled, Python 3.13+, and the required Python packages installed.
Store the username and password in a secure location and ensure that the Freqtrade API is not exposed publicly.
You would need to configure a separate MCP server instance for each Freqtrade instance, each with its own configuration.
The MCP server provides tools to fetch market data and bot configuration, which can be used for backtesting, but the backtesting logic needs to be implemented separately.
Use the `fetch_performance` tool to retrieve performance metrics and integrate them into a monitoring dashboard.
The MCP server will likely return an error. Implement error handling in your AI agent to handle such scenarios gracefully.
Refer to the Freqtrade documentation for information on rate limiting. Implement appropriate delays in your AI agent to avoid exceeding the limits.