Rasa Documentation

https://rasa.com/docs
Admin
The Rasa Platform helps teams build, test, deploy, and analyze AI agents at scale, offering pro-code...
Tokens:939,409
Snippets:7,614
Trust Score:10
Update:2 weeks ago
Show doc for...
Tokens:
Raw
### Run setup-assistant.sh for Rasa assistant

Source: https://rasa.com/docs/learn/deployment/gcp/gcp-playbook-rasa

Executes the initial setup for the Rasa assistant using the Rasa Helm Chart. This script is part of the GCP playbook and should be run after Kafka is installed.

```bash
./gcp/rasa/assistant/setup-assistant.sh

```

--------------------------------

### Example prompt to run all evaluation scenarios

Source: https://rasa.com/docs/pro/installation/rasa-mcp-tools

Example prompt to run all evaluation scenarios and get a summary of failures.

```plaintext
Run all scenarios in eval/scenarios/ and give me a summary of what failed.
```

--------------------------------

### Sample knowledge base document (docs/sampleDocument.txt)

Source: https://rasa.com/docs/learn/guides/integrating-rag

Sample knowledge base document stored in the /docs folder for folder-based storage. Used in the quickstart example to demonstrate how to prepare your knowledge base for Enterprise Search.

```text
We have five type of cards:  debit card, credit card, and loyalty card.

The price of a debit card is between 5 euros and 10, depending on the options you want to have.
Debit cards are directly linked to the user's bank account.

If the borrowed amount is not paid in full, credit card users may incur interest charges.

Loyalty cards are designed to reward customers for their repeat business.
Customers typically earn points or receive discounts based on their purchases.
Loyalty cards are often part of broader membership programs that may include additional benefits.

```

--------------------------------

### Install Rasa Pro with Helm

Source: https://rasa.com/docs/pro/deploy/deploy-kubernetes

Deploy Rasa Pro with Helm using the provided values file. This starts a Rasa Pro pod.

```bash
helm install \
    --namespace <your-namespace> \
    --values values.yml \
    <release-name> \
    rasa-<version>.tgz

```

--------------------------------

### Install uv on macOS and Linux

Source: https://rasa.com/docs/pro/installation/python

Installs the uv package manager using the official install script. Other installation methods are documented in the uv documentation.

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh

```

--------------------------------

### Create a project and install Rasa Pro

Source: https://rasa.com/docs/learn/quickstart/pro

Commands to create a new directory, set up a Python 3.11 virtual environment with uv, install rasa-pro, and initialize a basic Rasa project.

```bash
mkdir rasa-agent
cd rasa-agent
uv venv --python 3.11
source .venv/bin/activate
uv pip install rasa-pro
rasa init --template=basic

```

--------------------------------

### Retrieve all conversations for a user - request examples

Source: https://rasa.com/docs/reference/api/pro/http-api/tracker/retrieve-all-conversations-for-a-user

Example requests for the GET /users/{user_id}/trackers endpoint. All variants call the same URL with an empty token query parameter and require admin authentication. The cURL example uses the --request GET flag; the Python example uses the requests library; the JavaScript example uses fetch; the PHP example uses cURL; the Go example uses net/http; the Java example uses Unirest; the Ruby example uses net/http.

```bash
curl --request GET \
  --url 'http://localhost:5005/users/{user_id}/trackers?token='
```

```python
import requests

url = "http://localhost:5005/users/{user_id}/trackers?token="

response = requests.get(url)

print(response.text)
```

```javascript
const options = {method: 'GET'};

fetch('http://localhost:5005/users/{user_id}/trackers?token=', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_PORT => "5005",
  CURLOPT_URL => "http://localhost:5005/users/{user_id}/trackers?token=",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

```go
package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "http://localhost:5005/users/{user_id}/trackers?token="

	req, _ := http.NewRequest("GET", url, nil)

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

```java
HttpResponse<String> response = Unirest.get("http://localhost:5005/users/{user_id}/trackers?token=")
  .asString();
```

```ruby
require 'uri'
require 'net/http'

url = URI("http://localhost:5005/users/{user_id}/trackers?token=")

http = Net::HTTP.new(url.host, url.port)

request = Net::HTTP::Get.new(url)

response = http.request(request)
puts response.read_body
```

--------------------------------

### Initialize Rasa project with tutorial template

Source: https://rasa.com/docs/pro/tutorial

Initializes a new Rasa project using the tutorial template. Run this in an empty directory to set up the project structure for the tutorial.

```bash
rasa init --template tutorial
```

--------------------------------

### Retrieve a conversation's tracker with GET request

Source: https://rasa.com/docs/reference/api/pro/http-api/tracker/retrieve-a-conversations-tracker

Send a GET request to the tracker endpoint with the query parameters include_events=AFTER_RESTART, until=None, and an empty token. The Go example uses the standard net/http library, the Java example uses Unirest, and the Ruby example uses net/http.

```go
package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "http://localhost:5005/conversations/{conversation_id}/tracker?include_events=AFTER_RESTART&until=None&token="

	req, _ := http.NewRequest("GET", url, nil)

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

```java
HttpResponse<String> response = Unirest.get("http://localhost:5005/conversations/{conversation_id}/tracker?include_events=AFTER_RESTART&until=None&token=")
  .asString();
```

```ruby
require 'uri'
require 'net/http'

url = URI("http://localhost:5005/conversations/{conversation_id}/tracker?include_events=AFTER_RESTART&until=None&token=")

http = Net::HTTP.new(url.host, url.port)

request = Net::HTTP::Get.new(url)

response = http.request(request)
puts response.read_body
```

--------------------------------

### Setup database with setup-db.sh

Source: https://rasa.com/docs/learn/deployment/aws/aws-playbook-k8s

Sets up users and permissions on the database so that Rasa products can authenticate and use it.

```bash
./aws/configure/setup-db.sh

```

--------------------------------

### Initialize Rasa MCP Tools

Source: https://rasa.com/docs/learn/quickstart/pro

Runs the setup wizard that creates .rasa/tools.yaml and downloads offline docs and skills for your project.

```bash
rasa tools init

```

--------------------------------

### GET /status request examples in multiple languages

Source: https://rasa.com/docs/reference/api/pro/http-api/server-information/status-of-the-rasa-server

Send a GET request to the /status endpoint to check the status of the Rasa server. The token query parameter is required for authentication. The examples show how to make the request in various languages.

```cURL
curl --request GET \
  --url 'http://localhost:5005/status?token='
```

```Python
import requests

url = "http://localhost:5005/status?token="

response = requests.get(url)

print(response.text)
```

```JavaScript
const options = {method: 'GET'};

fetch('http://localhost:5005/status?token=', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

```PHP
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_PORT => "5005",
  CURLOPT_URL => "http://localhost:5005/status?token=",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

```Go
package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "http://localhost:5005/status?token="

	req, _ := http.NewRequest("GET", url, nil)

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

```Java
HttpResponse<String> response = Unirest.get("http://localhost:5005/status?token=")
  .asString();
```

```Ruby
require 'uri'
require 'net/http'

url = URI("http://localhost:5005/status?token=")

http = Net::HTTP.new(url.host, url.port)

request = Net::HTTP::Get.new(url)

response = http.request(request)
puts response.read_body
```

--------------------------------

### Example prompt to test conversations

Source: https://rasa.com/docs/pro/installation/rasa-mcp-tools

Example prompt to test conversations with the assistant after starting the Rasa server, including checking if a specific flow completed.

```plaintext
Talk to the assistant: "hello", "I want to send money to Jen", "$50", "yes".
Check if the transfer_money flow completed.
```

--------------------------------

### SessionStarted event constructor and example

Source: https://rasa.com/docs/reference/integrations/action-server/sdk-events

Defines the SessionStarted event to mark the start of a new session. The only parameter is an optional timestamp. The example creates an event with no arguments.

```python
SessionStarted(timestamp: Optional[float] = None)

```

```python
evt = SessionStarted()

```

--------------------------------

### Define training examples with simple format

Source: https://rasa.com/docs/reference/primitives/training-data-format

Shows the standard format for listing training examples one per line under an intent.

```yaml
nlu:
- intent: greet
  examples: |
    - hey
    - hi
    - whats up

```

--------------------------------

### Run setup-studio.sh to prepare Rasa Studio

Source: https://rasa.com/docs/learn/deployment/aws/aws-playbook-studio

Run this script to pull the Rasa Studio Helm chart, ensure secret values are set, and create a Kubernetes secret. It outputs the Keycloak admin credentials at the end, which are needed later for managing Studio users.

```bash
./aws/studio/setup-studio.sh

```

--------------------------------

### CRFEntityExtractor Output Example

Source: https://rasa.com/docs/reference/config/components/nlu-components

Example output of the CRFEntityExtractor component, showing the structure of extracted entities with value, start, end, entity type, confidence, and extractor name.

```json
{
  "entities": [
    {
      "value": "New York City",
      "start": 20,
      "end": 33,
      "entity": "city",
      "confidence": 0.874,
      "extractor": "CRFEntityExtractor"
    }
  ]
}
```

--------------------------------

### DucklingEntityExtractor Output Example

Source: https://rasa.com/docs/reference/config/components/nlu-components

Example output of the DucklingEntityExtractor component. It shows an extracted time entity with start and end positions, a normalized value, a confidence of 1.0, and the extractor name.

```json
{
  "entities": [
    {
      "end": 53,
      "entity": "time",
      "start": 48,
      "value": "2017-04-10T00:00:00.000+02:00",
      "confidence": 1.0,
      "extractor": "DucklingEntityExtractor"
    }
  ]
}
```

--------------------------------

### Set up database users and permissions

Source: https://rasa.com/docs/learn/deployment/azure/azure-playbook-k8s

Run this script to create database users and permissions for Rasa products.

```bash
./azure/configure/setup-db.sh

```

--------------------------------

### Start the action server with rasa run actions

Source: https://rasa.com/docs/rasa-pro/command-line-interface

Starts an action server with the Rasa SDK. Run 'rasa run actions --help' to see the full list of arguments.

```bash
rasa run actions

```

--------------------------------

### Example curl request to model server

Source: https://rasa.com/docs/rasa-pro/production/model-storage

Rasa sends a GET request to the configured URL with an If-None-Match header containing the hash of the last downloaded model. This example shows such a request.

```bash
curl --header "If-None-Match: d41d8cd98f00b204e9800998ecf8427e" http://my-server.com/models/default

```

--------------------------------

### Example domain.yml with session config, responses, slots, and actions

Source: https://rasa.com/docs/pro/build/configuring-assistant

Example domain.yml file defining the assistant's universe, including session configuration, responses, slots, and actions. The session_expiration_time is set to 60 minutes, and carry_over_slots_to_new_session is true.

```yaml
version: "3.1"

session_config:
  session_expiration_time: 60  # value in minutes, 0 means no timeout
  carry_over_slots_to_new_session: true

responses:
  utter_greeting:
    - text: "Hello! How can I help you today?"

slots:
  user_name:
    type: text
    initial_value: null

actions:
  - action_greet_user

```

--------------------------------

### Example prompt to validate a Rasa project

Source: https://rasa.com/docs/pro/installation/rasa-mcp-tools

A sample prompt to validate a Rasa project and get explanations of any issues.

```plaintext
Validate this Rasa project and explain any issues.
```

--------------------------------

### Run assistant setup script

Source: https://rasa.com/docs/learn/deployment/aws/aws-playbook-rasa

Runs the provided script to perform the initial setup for the Rasa assistant using the Rasa Helm Chart.

```bash
./aws/rasa/assistant/setup-assistant.sh

```

--------------------------------

### Install Rasa 3.7.0 Python package from new registry

Source: https://rasa.com/docs/rasa-pro/migration-guide

Starting from Rasa 3.7.0, the Python package is hosted on a new registry. Use the registry URL as an extra index URL when installing the rasa package with pip.

```text
https://europe-west3-python.pkg.dev/rasa-releases/rasa-plus-py

```

```bash
pip install  --extra-index-url=https://europe-west3-python.pkg.dev/rasa-releases/rasa-plus-py/simple rasa==3.7.0

```

--------------------------------

### Successful response body for GET /status

Source: https://rasa.com/docs/reference/api/pro/http-api/server-information/status-of-the-rasa-server

Example response body for a successful GET /status request. It includes the model ID, model file name, number of active training jobs, and the worker process ID.

```JSON
{
  "model_id": "75a985b7b86d442ca013d61ea4781b22",
  "model_file": "20190429-103105.tar.gz",
  "num_active_training_jobs": 2,
  "worker_pid": 12345
}
```

--------------------------------

### Local development connection string example

Source: https://rasa.com/docs/reference/integrations/analytics-database

Example connection string for local development. Uses default values: host localhost, port 5432, database studio, user studio_backend, and password studio_backend.

```text
postgresql://studio_backend:studio_backend@localhost:5432/studio
```

--------------------------------

### Go example for POST /model/parse

Source: https://rasa.com/docs/reference/api/pro/http-api/model/parse-a-message-using-the-rasa-model

Uses the net/http package to send a POST request with a JSON payload and prints the response body.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "http://localhost:5005/model/parse?token="

	payload := strings.NewReader("{\n  \"text\": \"Hello, I am Rasa!\",\n  \"message_id\": \"b2831e73-1407-4ba0-a861-0f30a42a2a5a\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

```

--------------------------------

### tests/e2e_test_cases.yml: Testing before first user utterance

Source: https://rasa.com/docs/reference/testing/e2e/test-cases

Example of testing the start of a conversation before the first user utterance. This is possible when action_session_start has been customized to also utter_welcome.

```yaml
test_cases:
  - test_case: user books a restaurant
    steps:
      - utter: utter_welcome # action_session_start is modified to also utter_welcome
      - user: I want to book a table for 4 at Xaigon for 8pm tonight
    ...
```

--------------------------------

### Go example

Source: https://rasa.com/docs/reference/api/pro/http-api/tracker/run-an-action-in-a-conversation

Uses the net/http package to send a POST request with a JSON payload to the execute endpoint. Prints the response body.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "http://localhost:5005/conversations/{conversation_id}/execute?include_events=AFTER_RESTART&token="

	payload := strings.NewReader("{\n  \"name\": \"utter_greet\",\n  \"policy\": \"<string>\",\n  \"confidence\": 0.987232\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

```

--------------------------------

### GET /conversations/{conversation_id}/tracker

Source: https://rasa.com/docs/rasa-pro/production/http-api

This snippet documents the endpoint for fetching a conversation tracker, which is explicitly mentioned in the source as an example of a request that requires authentication.

```APIDOC
## [GET] /conversations/{conversation_id}/tracker

### Description
Fetches the tracker for a given conversation. This endpoint requires authentication via a token or JWT.

### Method
GET

### Endpoint
/conversations/{conversation_id}/tracker

### Parameters
#### Path Parameters
- **conversation_id** (string) - Required - The ID of the conversation whose tracker is to be fetched.

#### Query Parameters
- **token** (string) - Required when using token-based authentication - The authentication token.

### Request Example
```
curl -XGET localhost:5005/conversations/default/tracker?token=thisismysecret
```

### Response
#### Success Response (200)
- **tracker** (object) - The tracker data for the conversation.

#### Response Example
```
{
  "tracker": {}
}
```

```
Tokens:
Raw