My Experience At Kestra Hackweek

I didn’t plan on pulling an all-nighter, but when things align perfectly,I couldn’t wait.

How It Started

I first heard about Kestra Hack Week through a notification in the WeMakeDevs Discord. I had Zero experience with Kestra, but the theme of the hack week was simple enough: to build something useful using their orchestration engine.

So I signed up, followed the onboarding, and got to work. I had no idea it would end with me winning the GitHub Star Award, but that’s how it played out :D

Getting Started with Kestra

I started by diving into Kestra’s Quickstart guide which is arguably one of the fastest I’ve ever tried. Running it on my Ubuntu machine only took a few Docker commands, and I had it up and running locally in minutes. Everything was so well-documented, and honestly, the onboarding was far better than I expected from a self-hosted tool.

For geting a feel for how flows worked, I explored:

• Official blueprints

• Git examples from the docs

• OpenAI task documentation

• Basic Structure Concepts like workspaces, triggers, secrets and variables

Within 2 to 3 hours, I had a good grip on how things in kestra worked.

Where did the Idea come from ?

This is where my open-source background kicked in.

Just for some context : I'm a maintainer at wasmCloud, and during one of our weekly community calls, we had a short discussion about developer experience. Someone on-call mentioned how useful it would be to have daily commit summaries, not raw logs, but something human-readable and contributor-friendly.

That idea really resonated with me because as someone who's been contributing to open source for a while, it's frustrating to pour time into a PR only to find out two other people were already working on the same thing. Your work ends up in the bin, and it’s no one’s fault, just a lack of visibility.

And then, literally the next day, I saw the Kestra Hack Week announcement.

The timing was literally unbelivable. I already had a pretty good real problem in mind. And I also had the perfect tool to build a solution.

About ChangelogGPT-flow

The idea was simple: automate the “What’s New” section of a project.

This is what the flow does:

1. Runs daily on a cron schedule

2. Clones the repository and pulls the last 24 hours of commits

3. Sends them to GPT-4 to generate a readable summary

4. Updates the changelog file automatically

I used a few existing blueprints to understand how to work with Git repos and API calls. I fine-tuned the OpenAI prompt to avoid generic output and to focus on clarity and structure.

The end result was a one-click-deployable flow that just works, especially for active repos where changelogs generally tend to lag behind the actual development cycle.

Winning the GitHub Star Award

I didn’t expect to win, but the response was quite strong. The project resonated with people because it solved a real, annoying problem and it did it cleanly.

Getting recognized with the GitHub Star Award was a nice bonus. But the real value for me was building something that I’m actually using and others can do too.

Final Thoughts

Kestra is powerful. I had zero experience with it, and within hours I had a fully working automation pipeline wired up with GPT-4 and GitHub.

If you’re maintaining a repo, contributing to fast-moving projects, or just want to automate the boring stuff, Kestra is definitely worth exploring.

Sometimes a random Discord ping ends up unlocking an entire workflow.

---

Links

• ChangelogGPT-flow on GitHub (it got renamed in the blueprints)

• Kestra

• wasmCloud