Inspiration
We really made this from scratch after the idea of reinventing a Kanban board.
What it does
Creates and organizes GitHub issues to a user specified and authorized repository.
How we built it
On entry the user is prompted to authorize this website to your GitHub account, then you select your owner and repo you would like to create issues into. The user authentication is saved for after the user enters all the steps (goals, constraints, context, guardrails) and selects generate. When generate is selected the authenticated details, and the input request payload is sent to a supabase edge function which is digested through a Gemini model to create and organize issues. These issues once completed are created into the authenticated repo.
Challenges we ran into
Authenticating using GitHub and storing tog credentials on Supabase to access github access_tokens to create their issues. Connecting and creating the rules via Gemini SDK.
Accomplishments that we're proud of
Fully complete authenticated and integrated tool.
What we learned
How to utilize AI models such as Gemini in a project. How to integrate OAuth app authorization from another application onto owned project.
What's next for ReKanban
If we were to continue working on this project as it is definitely in a prototype state we would allow users to use their own models on their own accounts through even different AI companies such as OpenAI and Claude. Other options we would want to include is a way for users to choose a Kanban board service such as Jira for the generated tasks to be created. Additionally we would want to refine how the AI generates the tasks including allowing different document formats instead of having to input all the data within our form. Some of the features we added came from resources within the Hackathon staff and greatly appreciate it as it pushed us to strive for a better version of our project.
Log in or sign up for Devpost to join the conversation.