Sure, everyone seeks happiness. But what kind of happiness do people want? At the end of story, we want the user to discuss if it’s happy ending, and think about what happiness means to you.

What it does

  • You can listen to various short stories of a girl who sells the matches.
  • A story offers you two choices that affect the fate.
  • After the ending, you can see the statistical data which show what percentage of people chose the each choice.
  • A story is about 5 minutes.
  • The skill plays a story randomly.
  • When you select another choice second time, you may discover something.


One day, a girl meets a street musician playing on the street. The audience seemed happy, but the police interrupted an unpermitted performance.

The musician ran away and the police asked the girl where he went. Which action do you take?

choice1. Tell the right direction

choice2. Tell the wrong direction

How I built it

We organized in a chart all the flows that users follow. a chart image

Challenges I ran into

  1. Limit number of audio files up to 5
  2. Preservation of statistical information on options selected by the user
  3. Screen transition at appropriate timing
  4. Structuring sentences for each story


  1. Combine multiple audio files into one file
  2. Use ask-sdk-s3-persistence-adapter library
  3. Designed to increase screen switching by using videos.
  4. Described in a JSON file so that it can be identified for each local sentence appearing on the story, voices of characters, audio files, etc..

You can also find details in this article in Japanese: link

What I learned

Making better VUI design. Unlike GUI, a user behavior cannot be restricted on the screen, so it is necessary to devise a way to ensure that the user can return to the normal flow anytime.

What's next for

  • Multi-language support
  • Add stories (the skill has 10 stories now)
  • Enhance voice actors’ performance
  • Add illustrations & scenes to all stories
  • Switch screens in the choiced phase
Share this project: