What is Deep Rewrite? How the Pro Rewrite Mode Works
Every rewrite in Deepclario makes your prompt better. Deep Rewrite makes it better twice: a stronger model writes the improved prompt, tears its own draft apart, and rewrites it before you ever see it.
This guide explains exactly what happens in each mode, what you get for the upgrade, and when each one is the right choice.
The short version
The normal rewrite gives you the model's first draft. Deep Rewrite makes a stronger model write it, critique it, and rewrite it. First drafts fix the obvious. The critique pass catches what first drafts miss.
The three passes
Draft
The model writes the full improved prompt the way the normal rewrite would: a clear role, one unambiguous task, format, constraints, and the context you gave it.
Critique
Then it attacks its own draft the way a strict senior reviewer would. Where is it still generic? Which element is weakest? What would ChatGPT most likely get wrong when given this prompt? Is any sentence filler?
Refine
Finally it rewrites the draft resolving every critique: a sharper role, tighter constraints, and a concrete example or acceptance criterion when the task benefits from one. Only this refined version is returned to you.
Normal rewrite vs Deep Rewrite
Normal
Deep Rewrite
Model
Fast, lightweight model
Stronger model that reasons more carefully and follows instructions better
Process
One pass: the first draft is the answer
Three passes: draft, critique, refine. You only see the refined version
What it catches
The obvious: missing role, unclear goal, no structure
The non-obvious: contradictory constraints, missing acceptance criteria, vague nouns, unstated assumptions
Rewrite length
Typically 80 to 250 words
Up to about 350 words when the task genuinely needs the detail
Speed
Fastest
A little slower: the extra passes are the point
Availability
Free and Pro
Pro only
Everything else stays the same in both modes: the five-dimension clarity score, the single clarifying question when something critical is missing, and the before/after score so you can see what changed.
What the critique pass actually catches
Take a prompt like this:
“Write an email to my team about the new process changes. Some people won't like it. Make it sound good but also firm.”
A first-draft rewrite fixes the surface: it adds a role, asks for a subject line, structures the email. But “sound good but also firm” is a contradiction the first draft usually just echoes back.
The critique pass flags it. The refined prompt resolves it into something the AI can actually execute: “empathetic in the opening, unambiguous about the decision being final, with one concrete channel for feedback.” That resolution is the difference between an email you send and an email you rewrite yourself.
When to use which
Normal rewrite
Quick one-off questions, brainstorming, everyday prompts where a solid improvement fast is worth more than a perfect one.
Deep Rewrite
Prompts where the output matters: specs, briefs, code-generation prompts, important emails, and any prompt you plan to reuse.
How to turn it on
In the playground, click the “Deep Rewrite” toggle next to the tone picker before you hit Improve. It stays on through any clarifying question. Deep Rewrite is part of the Pro plan, alongside unlimited rewrites, full history, and the weekly insights report.
Try it on your own prompt
Run the same prompt through the normal rewrite first, then with Deep Rewrite on, and compare the two side by side. The difference is easiest to see on a prompt you actually care about.