Show HN: Vizier – Codifying My Agent Development Workflow
JTan2231 Wednesday, November 26, 2025
I was getting tired of forgetting things and typing other things repeatedly into Codex's prompt, so I've been building some config + prompt glue to help with what I find to be the more tedious parts of development.
It's a little dirty and rough (I've taken some liberty in testing it on its own master branch), but I would love to know if this maps to anything close to a solution for your guys' aches with using agents. I always hear a million different experiences people have for the same problem with them, so I'm wondering if something like this would be a step toward reducing that number.
Summary
Vizier is an open-source deep learning-based image-to-image translation framework that allows users to perform various computer vision tasks, such as style transfer, semantic segmentation, and image restoration. The framework provides a modular design, enabling easy customization and integration of different models and datasets.
1
0
Summary
github.com