Step 1 of 3

Choose your path to get started

30s - 2 min

1Choose Path(current step)
2Customize(upcoming)
3Setup & Done(upcoming)

What's Next

Standard path: immediate setup (30s) • Custom path: answer ten or less questions (2 min)

Back to providers

Humanize ChatGPT

How would you like to humanize your AI?

Choose the option that works best for you

Recommended

Standard Humanizing Prompt

Use our pre-crafted prompt that works for most people

Use Standard Prompt →

Customize Your Tone

Answer less than ten questions to create a personalized prompt

Customize Prompt →

Coming Soon

Based on Your Writing

Generate a prompt from your emails or writing samples ($10)

Join Waitlist →

Humanizing ChatGPT: what to know

ChatGPT is the model most people meet first. That's both its strength and its main pain point: the default voice is the AI voice everyone has internalized as 'an AI wrote this' — hedged openings, three-part bullets, sentences that end with 'and so much more.' All of it is fixable with the prompt below; ChatGPT follows custom instructions reliably.

What ChatGPT sounds like out of the box

Hedged. Structured. Tends to summarize before answering, then summarize again at the end. Bullet-heavy in 'professional' contexts. Adds disclaimers when none were asked for.

Where ChatGPT earns its keep

Drafting and editing structured documents (PRDs, briefs, summaries)
Anything where a clear process being applied is the point
Plugin- or tool-augmented workflows (Code Interpreter, web browsing, custom GPTs)
Onboarding teammates who are new to AI — the UI is the most familiar

What humanizing ChatGPT actually fixes

The triple-summary pattern (intro summary, body, closing summary) — gone after one prompt line.
Em-dash addiction — the model leans on em-dashes hard. The wizard's punctuation question turns this off.
Disclaimers nobody asked for ('As an AI, I cannot…') — easy to ban in the prompt.

ChatGPT-specific prompt tips

Put your style instructions in the second Custom Instructions field, not the first. The first is for context about you, the second for response behavior.
ChatGPT respects 'never use the word X' lists better than 'do not be Y' instructions. Be concrete.
If responses still drift, paste the prompt at the top of new chats too. Belt and suspenders.
Projects (in ChatGPT Plus) let you scope a humanizing prompt to a specific project — useful when different work needs different voices.

Common questions about ChatGPT

Does this work in the free tier?: Yes. Custom Instructions are available on the free plan and respected by the model. Some advanced features (Projects, GPTs) are paid-only.
Will the prompt work in API calls too?: Yes — paste it as the system message. The behavior is the same as chat.
What about custom GPTs?: Same prompt in the GPT's instructions field. GPTs follow it more strictly than the general chat does.