Does this prompt work with Claude or Gemini?

Yes. No ChatGPT-specific feature is needed: paste it as-is into Claude, Gemini, or Le Chat and both versions come back the same way.

Why does the prompt return two versions instead of one?

Because "fix the tone" has no single right answer. Version A edits minimally, version B rewrites freely. You pick the one that sounds like you rather than accepting the first take.

Can the rewritten message distort what I meant?

The rule forbids it: every factual point must be kept. The prompt changes register, length, and tone, never the substance. If a fact was missing from your message, it will not be added.

← All prompts

Meetings & Communication · Updated July 17, 2026

ChatGPT prompt to rewrite a message and fix its tone

This is the prompt we run when a message is too blunt, too long, or simply off-brand. Its core rule: it adjusts tone and length while keeping every factual point, and it hands you a choice between two versions.

The prompt to copy

Rewrite this message to be {warmer / firmer / shorter / clearer} while keeping every factual point. Give me two versions: (A) minimal edit, (B) rewritten freely. Message: {paste}

Replace the field in curly braces with the adjustment you want, then paste the message as-is.

When to use it

On any message you hesitate to send: a reply written too fast and too hard, an email that drags, a nudge whose tone does not match the relationship. Also useful before sending something important to a client or a manager, when you want to compare two registers before deciding.

What you get back

Version A: your message, minimally edited.
Version B: a freer rewrite, more committed.
The same facts in both, only the tone changes.
A real choice, instead of one take to accept or reject.

Real example: a blunt message → two versions

Input: a message Tom, at Brightside, was about to send to Fernhill Coffee. Requested adjustment: warmer. The message, as-is:

Hi, we haven't received your brand guidelines. Without them we can't move forward on packaging. Please send them.

Output, the part that matters:

(A) minimal edit
Hi, a quick reminder: we're waiting on your brand guidelines to move forward on packaging. Could you send them over when you get a moment? Thanks so much.

(B) rewritten freely
Hi, hope all is well on your end. To kick off the packaging, we're just missing your brand guidelines. As soon as you forward them, we'll get going. Thanks in advance!

Notice what the AI did not do: it added no deadline and no reproach that the original message did not contain. It warmed the tone, nothing more. The core fact, that the missing brand guidelines are blocking packaging, appears identically in both versions.

The full use cases, with several registers and before/after examples, ship inside the kit.

Frequently asked questions

Does this prompt work with Claude or Gemini?: Yes. No ChatGPT-specific feature is needed: paste it as-is into Claude, Gemini, or Le Chat and both versions come back the same way.
Why does the prompt return two versions instead of one?: Because "fix the tone" has no single right answer. Version A edits minimally, version B rewrites freely. You pick the one that sounds like you rather than accepting the first take.
Can the rewritten message distort what I meant?: The rule forbids it: every factual point must be kept. The prompt changes register, length, and tone, never the substance. If a fact was missing from your message, it will not be added.