Program the Amazon Echo with Python

Flask-Ask is a Flask extension that makes building Alexa skills for the Amazon Echo easier and much more fun.

The Basics

A Flask-Ask application looks like this:

from flask import Flask
from flask_ask import Ask, statement

app = Flask(__name__)
ask = Ask(app, '/')

def hello(firstname):
    speech_text = "Hello %s" % firstname
    return statement(speech_text).simple_card('Hello', speech_text)

if __name__ == '__main__':

In the code above:

  1. The Ask object is created by passing in the Flask application and a route to forward Alexa requests to.
  2. The intent decorator maps HelloIntent to a view function hello.
  3. The intent’s firstname slot is implicitly mapped to hello‘s firstname parameter.
  4. Lastly, a builder constructs a spoken response and displays a contextual card in the Alexa smartphone/tablet app.

More code examples are in the samples directory.

Jinja Templates

Since Alexa responses are usually short phrases, you might find it convenient to put them in the same file. Flask-Ask has a Jinja template loader that loads multiple templates from a single YAML file. For example, here’s a template that supports the minimal voice interface above:

hello: Hello, {{ firstname }}

Templates are stored in a file called templates.yaml located in the application root. Checkout the Tidepooler example to see why it makes sense to extract speech out of the code and into templates as the number of spoken phrases grow.


Flask-Ask handles the boilerplate, so you can focus on writing clean code. Flask-Ask:

  • Has decorators to map Alexa requests and intent slots to view functions
  • Helps construct ask and tell responses, reprompts and cards
  • Makes session management easy
  • Allows for the separation of code and speech through Jinja templates
  • Verifies Alexa request signatures


To install Flask-Ask:

pip install flask-ask


These resources will get you up and running quickly:

Fantastic 3-part tutorial series by Harrison Kinsley


You can deploy using any WSGI compliant framework (uWSGI, Gunicorn). If you haven’t deployed a Flask app to production, checkout flask-live-starter.

To deploy on AWS Lambda, you have two options. Use Zappa to automate the deployment of an AWS Lambda function and an AWS API Gateway to provide a public facing endpoint for your Lambda function. This blog post shows how to deploy Flask-Ask with Zappa from scratch. Note: When deploying to AWS Lambda with Zappa, make sure you point the Alexa skill to the HTTPS API gateway that Zappa creates, not the Lambda function’s ARN.

Alternatively, you can use AWS Lambda directly without the need for an AWS API Gateway endpoint. In this case you will need to deploy your Lambda function yourself and use virtualenv to create a deployment package that contains your Flask-Ask application along with its dependencies, which can be uploaded to Lambda. If your Lambda handler is configured as lambda_function.lambda_handler, then you would save the full application example above in a file called and add the following two lines to it:

def lambda_handler(event, _context):
    return ask.run_aws_lambda(event)


If you’d like to work from the Flask-Ask source, clone the project and run:

pip install -r requirements-dev.txt

This will install all base requirements from requirements.txt as well as requirements needed for running tests from the tests directory.

Tests can be run with:

python test


python -m unittest

To install from your local clone or fork of the project, run:

python install

Related projects

cookiecutter-flask-ask is a Cookiecutter to easily bootstrap a Flask-Ask project, including documentation, speech assets and basic built-in intents.

Have a Google Home? Checkout Flask-Assistant (early alpha)

Thank You

Thanks for checking this library out! I hope you find it useful.

Of course, there’s always room for improvement. Feel free to open an issue so we can make Flask-Ask better.

Special thanks to @kennethreitz for his sense of style, and of course, @mitsuhiko for Flask