Launching today

EchoRun

Launching today

A runner where your past habits become your next opponent

0 followers

A runner where your past habits become your next opponent

0 followers

Visit website

EchoRun is an open-source 3D runner where your past playstyle becomes your next opponent. Built largely with GPT-6 Astra, including scene creation, 3D modeling, and gameplay code, it challenges you to break your habits and outsmart your previous self. Play in your browser or explore the MIT-licensed source on GitHub.

Free

Launch tags:Open Source•Artificial Intelligence•GitHub

Launch Team / Built With

OpenAI

GPT-6 Astra Challenge

How did Astra change the scope or ambition of what you built?

Maker

GPT-6 Astra helped me take on a much broader scope as a solo developer. Rather than focusing only on gameplay code, I could also work on scene creation and 3D modeling within the same project. Astra helped build most of EchoRun, while I guided the creative direction and iterated on the results. That made it possible to pursue a playable 3D browser game built around an ambitious idea: your past playstyle becomes your next opponent, challenging you to recognize and break your own habits. Its contribution went beyond individual code snippets to helping connect the game's environment, visuals, and mechanics into a playable experience. The source is available under the MIT license. Astra assisted development; the adaptive opponent runs locally during play.

Report

5h ago

Maker

📌

Hi Product Hunt! I'm the creator of EchoRun. What if your next opponent learned from the way you played? That idea became EchoRun, a 3D runner where your past playstyle shapes a shadow opponent. To beat it, you need to recognize your habits and change your strategy. GPT-6 Astra helped build most of the project, including scene creation, 3D modeling, and gameplay code. I guided the creative direction, played through the results, and iterated with Astra on what worked and what needed improvement. Astra's role was in development; the game's adaptive opponent runs locally, without calling a hosted language model during play. You can try the demo in your browser and explore the MIT-licensed source on GitHub. I'd love to hear: can you tell what the shadow has learned, and does it make you want to try a different strategy? Thanks for playing!

Report

1d ago