About Lucid

Hi, I'm Stavros. I built Lucid because I love writing, and I was worried AI was going to take that away from me.

A couple of years ago I started using LLMs to help with my drafts, and at first it felt like a superpower. Type a rough idea, get back a polished paragraph. But after a while the paragraphs started to bother me. They didn't sound like me. They sounded like an AI doing its impression of a writer, which is a particular voice, and not one I wanted to publish under my name.

So I'd spend twenty minutes editing the AI's paragraph to sound less like the AI. Then ten more minutes wondering if the original idea had even survived. By the end I'd written less than if I'd just sat down and written the thing myself, and I felt worse about the result. Something had gone wrong.

The annoying part is that AI is useful for writers. It catches things I miss. It notices when I've used the same word three times in a paragraph, when a transition is doing too much work, when an argument has a hole in it. That kind of feedback is gold, and most of us don't have an editor on call to give it to us.

The problem is the easiest way to deliver that feedback with today's AI tools is to just rewrite the thing. Which makes the editor also the ghostwriter, and the line between "this could be tightened" and "let me tighten it for you" gets blurry fast.

Lucid draws that line back. You write, the AI critiques. Every suggestion lives in a margin comment attached to a specific sentence, like notes from a copyeditor in the margin of a manuscript. You can revise based on it, reply if you disagree, or archive it and move on. The AI never touches the prose. That part stays yours.

If you write essays, blog posts, fiction, reports, or anything long-form where voice matters, Lucid is for you. There's a free tier with three critiques a month, and you can bring your own Anthropic key for more.

Try it on a draft you're working on. No signup needed to start.

Questions? Email me. I read everything.