Rasa SDK

Rasa SDK provides the tools you need to write custom actions in python.


Use pip to install rasa-sdk on your action server.

pip install rasa-sdk


You do not need to install rasa for your action server. E.g. if you are running Rasa in a docker container, it is recommended to create a separate container for your action server. In this separate container, you only need to install rasa-sdk.

Running the Action Server

If you have rasa installed, run this command to start your action server:

rasa run actions

Otherwise, if you do not have rasa installed, run this command:

python -m rasa_sdk --actions actions

You can verify that the action server is up and running with the command:

curl http://localhost:5055/health

You can get the list of registered custom actions with the command:

curl http://localhost:5055/actions

The file that contains your custom actions should be called actions.py. Alternatively, you can use a package directory called actions or else manually specify an actions module or package with the --actions flag.

The full list of options for running the action server with either command is:

usage: rasa run actions [-h] [-v] [-vv] [--quiet] [-p PORT]
                        [--cors [CORS [CORS ...]]] [--actions ACTIONS]
                        [--ssl-keyfile SSL_KEYFILE]
                        [--ssl-certificate SSL_CERTIFICATE]
                        [--ssl-password SSL_PASSWORD] [--auto-reload]

optional arguments:
  -h, --help            show this help message and exit
  -p PORT, --port PORT  port to run the server at (default: 5055)
  --cors [CORS [CORS ...]]
                        enable CORS for the passed origin. Use * to whitelist
                        all origins (default: None)
  --actions ACTIONS     name of action package to be loaded (default: None)
  --ssl-keyfile SSL_KEYFILE
                        Set the SSL certificate to create a TLS secured
                        server. (default: None)
  --ssl-certificate SSL_CERTIFICATE
                        Set the SSL certificate to create a TLS secured
                        server. (default: None)
  --ssl-password SSL_PASSWORD
                        If your ssl-keyfile is protected by a password, you
                        can specify it using this paramer. (default: None)
  --auto-reload         Enable auto-reloading of modules containing Action
                        subclasses. (default: False)

Python Logging Options:
  -v, --verbose         Be verbose. Sets logging level to INFO. (default:
  -vv, --debug          Print lots of debugging statements. Sets logging level
                        to DEBUG. (default: None)
  --quiet               Be quiet! Sets logging level to WARNING. (default:


The Action class is the base class for any custom action. It has two methods that both need to be overwritten, name() and run().

In a restaurant bot, if the user says “show me a Mexican restaurant”, your bot could execute the action ActionCheckRestaurants, which might look like this:

from rasa_sdk import Action
from rasa_sdk.events import SlotSet

class ActionCheckRestaurants(Action):
   def name(self) -> Text:
      return "action_check_restaurants"

   def run(self,
           dispatcher: CollectingDispatcher,
           tracker: Tracker,
           domain: Dict[Text, Any]) -> List[Dict[Text, Any]]:

      cuisine = tracker.get_slot('cuisine')
      q = "select * from restaurants where cuisine='{0}' limit 1".format(cuisine)
      result = db.query(q)

      return [SlotSet("matches", result if result is not None else [])]

You should add the action name action_check_restaurants to the actions in your domain file. The action’s run() method receives three arguments. You can access the values of slots and the latest message sent by the user using the tracker object, and you can send messages back to the user with the dispatcher object, by calling dispatcher.utter_message.

Details of the run() method:

async Action.run(dispatcher, tracker, domain)

Execute the side effects of this action.

  • dispatcher – the dispatcher which is used to send messages back to the user. Use dipatcher.utter_message() or any other rasa_sdk.executor.CollectingDispatcher method.

  • tracker – the state tracker for the current user. You can access slot values using tracker.get_slot(slot_name), the most recent user message is tracker.latest_message.text and any other rasa_sdk.Tracker property.

  • domain – the bot’s domain


A dictionary of rasa_sdk.events.Event instances that is

returned through the endpoint

Return type

List[Dict[str, Any]]

Details of the dispatcher.utter_message() method:

CollectingDispatcher.utter_message(text=None, image=None, json_message=None, template=None, attachment=None, buttons=None, elements=None, **kwargs)

“Send a text to the output channel

Return type


Customizing the session start action

The default behavior of the session start action is to take all existing slots and to carry them over into the next session. Let’s say you do not want to carry over all slots, but only a user’s name and their phone number. To do that, you’d override the action_session_start with a custom action that might look like this:

from typing import Text, List, Dict, Any

from rasa_sdk import Action, Tracker
from rasa_sdk.events import SlotSet, SessionStarted, ActionExecuted, EventType
from rasa_sdk.executor import CollectingDispatcher

class ActionSessionStart(Action):
    def name(self) -> Text:
        return "action_session_start"

    def fetch_slots(tracker: Tracker) -> List[EventType]:
        """Collect slots that contain the user's name and phone number."""

        slots = []

        for key in ("name", "phone_number"):
            value = tracker.get_slot(key)
            if value is not None:
                slots.append(SlotSet(key=key, value=value))

        return slots

    async def run(
        dispatcher: CollectingDispatcher,
        tracker: Tracker,
        domain: Dict[Text, Any],
    ) -> List[EventType]:

        # the session should begin with a `session_started` event
        events = [SessionStarted()]

        # any slots that should be carried over should come after the
        # `session_started` event

        # an `action_listen` should be added at the end as a user message follows

        return events


You need to explicitly add action_session_start to your domain to override this custom action.


An action’s run() method returns a list of events. For more information on the different types of events, see Events. There is an example of a SlotSet event above. The action itself will automatically be added to the tracker as an ActionExecuted event. If the action should not trigger any other events, it should return an empty list.


The rasa_sdk.Tracker lets you access the bot’s memory in your custom actions. You can get information about past events and the current state of the conversation through Tracker attributes and methods.

The following are available as attributes of a Tracker object:

  • sender_id - The unique ID of person talking to the bot.

  • slots - The list of slots that can be filled as defined in the “ref”domains.

  • latest_message - A dictionary containing the attributes of the latest message: intent, entities and text.

  • events - A list of all previous events.

  • active_form - The name of the currently active form.

  • latest_action_name - The name of the last action the bot executed.

The available methods from the Tracker are:


Return the current tracker state as an object.

Return type

Dict[str, Any]


State whether the tracker is currently paused.

Return type


Tracker.get_latest_entity_values(entity_type, entity_role=None, entity_group=None)

Get entity values found for the passed entity type and optional role and group in latest message.

If you are only interested in the first entity of a given type use next(tracker.get_latest_entity_values(“my_entity_name”), None). If no entity is found None is the default result.

  • entity_type – the entity type of interest

  • entity_role – optional entity role of interest

  • entity_group – optional entity group of interest


List of entity values.

Return type



Get the name of the input_channel of the latest UserUttered event

Return type



Return a list of events after the most recent restart.

Return type



Retrieves the value of a slot.

Return type