Actions

Action management system for conversation flows.

This module provides the ActionManager class which handles execution of actions during conversation state transitions. It supports:

• Built-in actions (TTS, conversation ending)

• Custom action registration

• Synchronous and asynchronous handlers

• Pre and post-transition actions

• Error handling and validation

Actions are used to perform side effects during conversations, such as:

• Text-to-speech output

• Database updates

• External API calls

• Custom integrations

class pipecat_flows.actions.FunctionActionFrame(action: dict, function: Callable[[dict[str, Any], FlowManager], Awaitable[None]])

Bases: ControlFrame

Frame containing a function action to be executed.

Parameters:

• action – Action configuration dictionary.

• function – Function handler to execute.

action: dict

function: Callable[[dict[str, Any], FlowManager], Awaitable[None]]

class pipecat_flows.actions.ActionFinishedFrame

Bases: ControlFrame

Frame indicating that an action has completed execution.

class pipecat_flows.actions.ActionManager(worker: PipelineWorker, flow_manager: FlowManager)

Bases: object

Manages the registration and execution of flow actions.

Actions are executed during state transitions and can include:

• Text-to-speech output

• Database updates

• External API calls

• Custom user-defined actions

Built-in actions:

• tts_say: Speak text using TTS

• end_conversation: End the current conversation

• function: Execute inline functions in the pipeline

Custom actions can be registered using register_action().

__init__(worker: PipelineWorker, flow_manager: FlowManager)

Initialize the action manager.

Parameters:

• worker – PipelineWorker instance used to queue frames.

• flow_manager – FlowManager instance that this ActionManager is part of.

async execute_actions(actions: list[ActionConfig] | None) → None

Execute a list of actions.

Parameters:

actions – List of action configurations to execute.

Raises:

ActionError – If action execution fails.

Note

Each action must have a ‘type’ field matching a registered handler.

schedule_deferred_post_actions(post_actions: list[ActionConfig]) → None

Schedule “deferred” post-actions to be executed after next LLM completion.

Parameters:

post_actions – List of actions to execute after LLM response.

clear_deferred_post_actions() → None

Clear any scheduled deferred post-actions.