New to the field as a junior dev, the goal is to set up a bigger idea now, as efficiently as possible and with a solid foundation, so it does not come back to bite later. The aim is to factor in future topics like "vibe coding" or "AI slop" as early as possible, and avoid problems right away. So the question is, how should a monorepo be structured sensibly, and what should be paid attention to from the very beginning?
Right now it feels like there are many things that should be considered, for example in an agent.md file and also generally as guiding principles.
The long term goal is an ecosystem of multiple products where users can register and log in everywhere with a single account.
Which concepts and frameworks make sense for this? Which AI driven concepts and clean architecture approaches would be recommended?
What are best practices, what is currently considered state of the art?
Is there an agent framework that is an obvious fit? GSD is something that has been mentioned, does it make sense to initialize the project with it right away?
What is currently on the list
Work is being done with Codex and Claude Code.
- Monorepo: Next.js + Expo + tRPC + Supabase
- Vector DB
- Kontext7 MCP
- GitHub Spec Kit
- DRY principle
- Atomic and feature component folder structure
Top comments (0)