freqtrade/freqtrade · error · OperationalException

You are trying to use a FreqAI strategy with process_only_ne

Error message

You are trying to use a FreqAI strategy with process_only_new_candles = False. This is not supported by FreqAI, and it is therefore aborting.

What it means

FreqAI's dry/live execution path (FreqAI.extract_data_and_train_model / update_scheduled_model) requires strategy.process_only_new_candles = True. FreqAI trains on a fixed candle snapshot, and processing every throttle iteration would retrain/repeat work inconsistently, so a False value aborts with OperationalException.

Source

Thrown at freqtrade/freqai/freqai_interface.py:447

        dk.append_predictions(append_df)
        dk.save_backtesting_prediction(append_df)

    def start_live(
        self, dataframe: DataFrame, metadata: dict, strategy: IStrategy, dk: FreqaiDataKitchen
    ) -> FreqaiDataKitchen:
        """
        The main broad execution for dry/live. This function will check if a retraining should be
        performed, and if so, retrain and reset the model.
        :param dataframe: DataFrame = strategy passed dataframe
        :param metadata: Dict = pair metadata
        :param strategy: IStrategy = currently employed strategy
        dk: FreqaiDataKitchen = Data management/analysis tool associated to present pair only
        :returns:
        dk: FreqaiDataKitchen = Data management/analysis tool associated to present pair only
        """

        if not strategy.process_only_new_candles:
            raise OperationalException(
                "You are trying to use a FreqAI strategy with "
                "process_only_new_candles = False. This is not supported "
                "by FreqAI, and it is therefore aborting."
            )

        # get the model metadata associated with the current pair
        (_, trained_timestamp) = self.dd.get_pair_dict_info(metadata["pair"])

        # append the historic data once per round
        if self.dd.historic_data:
            self.dd.update_historic_data(strategy, dk)
            logger.debug(f"Updating historic data on pair {metadata['pair']}")
            self.track_current_candle()

        (_, new_trained_timerange, data_load_timerange) = dk.check_if_new_training_required(
            trained_timestamp
        )
        dk.set_paths(metadata["pair"], new_trained_timerange.stopts)

View on GitHub (pinned to 1c8edfe4d1)

Solutions

Set process_only_new_candles = True in the FreqAI strategy class
Remove or set 'process_only_new_candles': true in the config JSON
Use the FreqaiExampleStrategy template, which ships with the correct value

Example fix

# before
class MyFreqaiStrategy(IStrategy):
    process_only_new_candles = False

# after
class MyFreqaiStrategy(IStrategy):
    process_only_new_candles = True

Defensive patterns

Strategy: validation

Validate before calling

# Guard before 'freqtrade trade'
process_only_new = config.get('process_only_new_candles',
                               getattr(strategy, 'process_only_new_candles', True))
assert process_only_new is True, 'FreqAI requires process_only_new_candles = True'

Type guard

def freqai_strategy_ready(strategy) -> bool:
    return bool(getattr(strategy, 'process_only_new_candles', False)) is True

Prevention

Set process_only_new_candles = True in every FreqAI strategy
Make sure the config does not override the strategy attribute to false

When it happens

Trigger: Starting dry-run or live trading ('freqtrade trade') with a FreqAI-enabled strategy and config where process_only_new_candles is set to False (in strategy attribute or config 'process_only_new_candles': false).

Common situations: User starts from a scalping strategy template that sets process_only_new_candles = False for faster reaction; config overrides the strategy attribute with 'process_only_new_candles': false.

Related errors

AI-assisted analysis of freqtrade/freqtrade@1c8edfe4d1 (2026-08-15). Data as JSON: /api/errors/4ad8377dabe25020. Report an issue: GitHub.