Non-developer building a production app with Claude : am I wasting usage with my workflow?
retained text
I’m a solo founder/non-developer building a fairly substantial PWA. I’m currently on Build 1.5.2 and the app is already deployed and being tested. My workflow has been: [Claude.ai](http://claude.ai/) → discuss feature/problem → have it create implementation instructions/code → copy that into Claude Code in VS Code → Claude Code edits the repo/tests it → take results back to [Claude.ai](http://claude.ai/) when needed. It worked extremely well while I was learning, but now it feels redundant. Claude Code has the actual repo, can inspect files, run tests/builds, debug, and make the changes itself. Meanwhile I’m burning a lot of Claude usage maintaining essentially two conversations about the same codebase. For other non-developers/solo founders who have gotten this far with Claude Code: **what does your actual workflow look like now?** Specifically, what do you keep in [Claude.ai](http://claude.ai/) vs Claude Code? How do you structure CLAUDE.md/project documentation? Do you plan and implement entirely inside Claude Code? And what Git/checkpoint workflow keeps Claude from breaking working features? I’m not trying to make Claude Code fully autonomous. I want to get better at supervising it without unnecessarily duplicating work between [Claude.ai](http://claude.ai/) and Claude Code.
document fetched 2026-09-18 00:41 · canonical https://reddit.com/r/ClaudeCode/comments/1wiu45g/nondeveloper_building_a_production_app_with