See example → BETA · FREE
✦ Free · No sign-up · AI-powered critique in 30 seconds

You know something's wrong
with this draft.
You just don't know what.

Paste any technical blog post, doc, or README. CritKit simulates three real reader types and quotes the exact sentence where each one stumbled.

Your 3 readers
🤔
The Confused Beginner
Flags every assumption you didn't realize you were making. The terms you defined in your head but not on the page. The steps you skipped because they felt obvious.
🧐
The Skeptical Expert
Spots the oversimplifications, the missing caveats, the places where you sacrificed accuracy for approachability. Won't let vague claims slide.
💼
The Potential Customer
Asks whether they can trust you, whether your claims are backed up, whether this piece gives them a reason to act. The hardest reader to satisfy.
critiques run · Free · No account needed
Your draft 0 words
Reading your draft...
🤔 Simulating the confused beginner
🧐 Consulting the skeptical expert
💼 Checking with the potential customer
📝 Compiling top actions
Critique for: your draft
Top 3 actions — start here

Ready to revise?

Make your changes and run it again — critique is most useful in rounds.

View example

How it works

1
Paste your draft
Blog post, docs page, README, landing copy — anything technical you want a second opinion on.
2
Claude reads it
The AI simulates three distinct readers, each with different background knowledge and goals.
3
Get specific quotes
Every issue includes a quoted sentence from your draft — so you know exactly where the problem is.
4
Revise and repeat
Critique is most useful in rounds. Make changes, paste again, watch the scores improve.

Why three readers?

Because one reader only ever finds one kind of problem.

🤔
The Confused Beginner
catches: gaps, missing context, assumed knowledge
Typical flag: "You used this term on step 2 without defining it. I had to Google it — and then I wasn't sure I was back in the right place."
🧐
The Skeptical Expert
catches: oversimplifications, missing caveats, inaccuracies
Typical flag: "This works in the happy path. You didn't mention what happens when the token expires mid-request — that's going to bite someone."
💼
The Potential Customer
catches: trust gaps, weak claims, missing reasons to act
Typical flag: "You said 'easy to set up' but the setup section is 14 steps. I'm not sure I believe the rest of your claims now."

Each reader has a completely different reason to give up on your draft. A single perspective — even a smart one — only ever catches one layer.