occupiedbyhim/prompt-engineer ? reverse-engineered prompt
Reverse engineered prompt
Build me a web app called Prompt Engineer that helps me compare different versions of the same prompt and prove which one actually works better in production.
I want to be able to create a prompt, save versions of it like drafts and commits, pin one version as the current production version, and run experiments where traffic is split between variants like 50/50 or 90/10. The app should send requests through either Anthropic or OpenAI, record the result, and show useful analytics like latency, tokens, cost, success rate, and whether the difference is statistically meaningful. I also want a simple way to mark an execution as successful later, after I know the outcome.
Make it feel polished and modern, with a dashboard for prompts, versions, experiments, and analytics. If you need to check current docs for the model APIs or anything else, go ahead and look them up online.
Are you gonna build this?
make sure you review the code using coderabbit