Inspiration

Our inspiration for Waggle came from large school projects and research, where multiple people would rattle off issues or bugs in quick succession, and we would have to circle back on these topics again and again, sometimes not catching everything, or mislabeling issues even after multiple rounds of the same topic. We wanted to have something which would capture every issue the moment it's spoken, allowing parallel work in finding and filing tickets for said issues.

What it does

Waggle is a TypeScript CLI which takes the full transcript from Bee, and automatically creates GitHub issues from all bugs, features, to-do, and other action item mentioned. Simply using your Bee wearable and talking fills the backlog automatically.

How we built it

Waggle was built with loop-engineering through an agentic system composed of Claude and Codex.

Challenges we ran into

Neither of us had used any Amazon-related technologies before, so learning how to interact with the Bee API from scratch, and connecting it through the MCP to build Waggle's pipeline.

Accomplishments that we're proud of

We went from never touching an API similar to Bee, to a working end-to-end pipeline: a conversation happens, and real GitHub issues show up in repo. Seeing issues pop up automatically after finishing our conversation was the moment this project felt real.

What we learned

We learned how to build on Bee and similar Amazon technologies, as well as how to integrate an MCP with it to turn raw conversational data to workable, structured output.

What's next for Waggle

The next steps for Waggle include: expanding to integrating multiple devices at once, increasing the bandwidth of technologies it supports, such as Jira or Linear, and linking related issues together for easier organization.

Built With

Share this project:

Updates

Submission history