Launched this week
PromptDev
PromptDev: Professional AI prompts with Ctrl+Q
7 followers
PromptDev: Professional AI prompts with Ctrl+Q
7 followers
PromptDev.site stands out by turning prompt engineering into actual software architecture. Unlike text-heavy prompt libraries, it provides a dedicated, developer-first sandbox to construct, stack, and real-time benchmark complex system instructions. It eliminates LLM guesswork with zero friction, making it the ultimate playground for full-stack developers building multi-agent setups or structured API integrations. Built for absolute speed, precision, and production-level workflow iteration.
Love the developer-first angle here, feels way more practical than the usual prompt dumps. One thing that would make it stick for me is a built-in diff view so I can see exactly what changed between prompt versions when I'm tweaking for a benchmark run. Right now I'm exporting and using another tool just to track what actually moved the needle, and having that natively would save a lot of context switching.
@bayrambalk7jpj Thanks for the awesome feedback, Bayram! I completely get your point. Context switching just to compare prompt iterations is a huge friction point for developers. Adding a native built-in diff view to track structural changes between version benchmarks is a fantastic idea and makes perfect sense for our workflow. I am adding this straight to the high-priority roadmap. Stay tuned, and thanks for helping make PromptDev better!
@bayrambalk7jpj Hey again! I'm back to let you know that I really liked your idea, and I've successfully implemented the comparison feature. It's live now, so feel free to check it out!
Genuinely impressed by how much control the stacking and benchmarking flow gives you, it actually feels closer to a real dev tool than another prompt playground.
@emirhana0qu Thank you so much, Emirhan! "Closer to a real dev tool than another prompt playground" is exactly what I was aiming for when building this. I wanted to treat prompts as dynamic system dependencies rather than just static text dumps. Really glad the stacking and benchmarking flow is giving you the level of control you need. Appreciate the support