Last month I had a prompt that finally clicked.
It was one of those rare ones: clear role, tight constraints, output that didn’t need three more rounds. I remember thinking “okay, this is the version — I should keep this.”
Then I didn’t.
The chat scrolled. I jumped to another model. A new tab ate the old one. A week later I needed the same shape of answer and… I rewrote it from memory. Worse. Slower. Slightly wrong in the ways you only notice after you’ve already shipped the draft.
That kept happening. Not because I don’t care about process — because the “save it properly” step always landed after the win, and after the win my brain had already moved on.
So I stopped pretending a notes app was going to catch me in that exact second.
I built PromptAlo for myself first: a free Chrome extension with one small job.
Add to Chrome → https://chromewebstore.google.com/detail/promptalo/jlfoipkmagdoafhlkhfkakdankpjhgni
When a prompt works, hit Save. It lands in a Library you can actually find. Reuse later — today that opens in Gemini. (ChatGPT / Claude one-click paste isn’t live yet. I’m not going to market it until it is.)
It’s not a second brain. It’s not a productivity religion. It’s just: stop losing the good ones.
I’m curious how other builders handle this — when a prompt finally works, where does it live a week later for you? And what would make you keep a tiny tool like this installed after week one?