Sketch the mess in plain words. Lucen forges it into a sharp, reusable prompt. Then save every one to a searchable library, so you never lose a good prompt again.

Which Vercel product did you use? What new capability did it give your team, and how did that shape the product you launched?

We run on Vercel's Next.js hosting with per-environment variables and Edge Middleware. That let us stand up a persistent dev deployment, gated by Basic-Auth and hidden from search, running against the same real Supabase and Stripe test-mode data as production. Our AI coding agents shipped every sprint straight to that environment and we verified real behavior, not mocks, before anything reached prod, without ever risking a staging URL leaking publicly pre-launch.

Hey everyone, I'm Akshita 👋 Lucen started with my own problem: I kept losing prompts that worked, and I spent too much time going back and forth with a model just to get something reusable. Here's how it works: you describe what you want in a sentence or two, Lucen asks a few clarifying questions to catch the gaps that actually matter (audience, constraints, edge cases), then forges a full system prompt in one of three modes: Quick, Balanced, or Deep. You can attach Tone, Format, Persona, or Guardrail blocks, and Lucen weaves them into the prompt as real rules rather than dropping your words in as-is. Everything gets saved to a searchable library so nothing good gets lost again. I'd love feedback from anyone who writes prompts regularly, where does Forge's output fall short, and does the extra step of clarifying questions feel worth it? Happy to answer anything else about the build too.

Which Vercel product did you use? What new capability did it give your team, and how did that shape the product you launched?

Lucen

Turn rough ideas into great prompts, saved in your library

Turn rough ideas into great prompts, saved in your library