AI-magine

• 

  Creating pictures for a famous novel starting from a user prompt

• 

  Illustrations can have different styles, such as illustrated or photographic

• 

  The experience is seamless for users, they just need a prompt and a style

Inspiration

We were inspired by seeing hundreds of people insert images into their confluence pages using Kolekti's Content Formatting Macro. We wanted to create a seamless experience for our macro users and highlight the potential of AI image generators, removing the hurdles of searching for and downloading images outside of confluence.

What it does

AI-magine is an AI-powered Macro that allows users to generate images for their confluence pages using a prompt and a style selector.

How we built it

We have built it using Forge/Typescript/React, hosted in a Bitbucket repo.

Challenges we ran into

One of the main limitations we ran into was Forge’s macro editor. Ideally, we would generate multiple images in the editor, which the user could choose between. Afterwards, users would select an image to be inserted into the confluence page. Unfortunately, this was not possible with the current macro editor components provided.

Accomplishments that we're proud of

We enjoyed the technical side of the project and getting to know Forge’s functionality and the possibilities within this new framework. We also had a lot of fun filming the movie presentation that accompanies our application.

What we learned

This project has been illuminating and given us insight into the best place to develop our application. In particular, we've learnt a lot about Forge, bitbucket and CI/CD.

What's next for AI-magine

We plan to develop the macro further and work towards releasing it on the Atlassian marketplace. We will always be on the lookout for new APIs that could be integrated into our application to improve its image capabilities.

Built With

• atlassian
• confluence
• dall-e
• forge
• hugging-face
• react
• stability-ai
• typescript