Most people use ChatGPT like they’re texting a friend. They send a vague request, give half instructions, and then follow up two or three more times to fix what went wrong. It turns into a back-and-forth that wastes time and still doesn’t produce exactly what they wanted.
Power users don’t work like that. Instead of writing long prompts every time, they use short code words at the top of their message to control how ChatGPT responds. These simple trigger words instantly set the tone, structure, and purpose of the output. It looks like this:
1. ROASTME
Use like:
ROASTME: paste your text
What it does:
Tears your writing apart like a savage editor. Funny, but still useful.
This is perfect for:
- Blog drafts
- Landing pages
- Cold emails
- Social media posts
It pushes ChatGPT to stop being polite and start being honest. You’ll usually get:
- Clear weaknesses
- Specific rewrite suggestions
- Overused phrases called out
- Logical gaps exposed
It’s like hiring a strict editor who actually wants your work to improve.
2. SPICY-TLDR
Use like:
SPICY-TLDR: paste your long text
What it does:
Summarizes your content with maximum sass and minimum words.
Then gives you a normal TLDR.
You get:
- A funny one-liner summary
- A clean, professional summary
This is great for:
- News breakdowns
- Reddit posts
- Explaining complex tech topics
- Internal docs
It forces clarity. If the “spicy” version makes no sense, your original probably needs tightening.
3. NPC-DIALOGUE
Use like:
NPC-DIALOGUE: explain this concept
What it does:
Rewrites your explanation like two awkward video-game NPCs talking.
It sounds silly. But it works.
Why? Because:
- It forces simple language
- It removes jargon
- It exposes confusion
If two NPCs can explain it clearly, your audience can understand it too.
4. CONSPIRACY-MODE
Use like:
CONSPIRACY-MODE: explain this topic
What it does:
Explains your topic like a conspiracy theorist…
Then immediately fact-checks itself and labels what’s real.
This is powerful for:
- Media literacy
- Breaking down viral claims
- Teaching critical thinking
- Analyzing hype around AI or tech
It shows how easy it is to distort facts, then corrects the distortion.
5. LAWYER-UP
Use like:
LAWYER-UP: rewrite this policy
What it does:
Rewrites content like a cautious lawyer wrote it.
Adds serious “this is not legal advice” energy.
Great for:
- Terms & conditions
- Disclaimers
- Policy pages
- Risk-sensitive content
It makes vague language precise and closes loopholes.
6. GOBLIN-SIMPLIFY
Use like:
GOBLIN-SIMPLIFY: explain this concept
What it does:
Explains something like a chaotic goblin…
But still keeps it correct and clear.
It’s surprisingly useful for:
- Technical topics
- Finance concepts
- AI explanations
- SEO tutorials
If the goblin can explain it simply, your blog readers can understand it too.
7. OVERCONFIDENT-INTERN
Use like:
OVERCONFIDENT-INTERN: rewrite this
What it does:
Rewrites content like an aggressively confident intern who’s slightly wrong.
Then adds a correction section.
This is excellent for:
- Spotting weak assumptions
- Finding hidden errors
- Stress-testing explanations
- Teaching through contrast
You see how misinformation can sneak in, then how to fix it.
8. HR-SAFETY
Use like:
HR-SAFETY: rewrite this announcement
What it does:
Sanitizes everything. Removes personality. Makes it corporate and neutral.
Then gives a human version.
Perfect for:
- Workplace emails
- Official announcements
- Internal communication
- Brand voice testing
You can instantly compare “corporate safe” vs “human and engaging.”
9. SITCOM-PITCH
Use like:
SITCOM-PITCH: explain this idea
What it does:
Turns your topic into a sitcom scene that teaches the concept.
Includes the lesson at the end.
This is fantastic for:
- Teaching marketing
- Explaining product strategy
- Breaking down AI concepts
- Writing educational content
Storytelling improves retention. Always.
10. BOSS-FIGHT
Use like:
BOSS-FIGHT: solve this problem
What it does:
Turns your problem into a game boss with:
- Phases
- Weaknesses
- Strategy guide
This works shockingly well for:
- Debugging
- Productivity challenges
- Business problems
- Learning roadmaps
It reframes overwhelm into structured action.
Why These Code Words Actually Work
This isn’t magic.
It’s structure.
Most users treat ChatGPT like a person.
Power users treat it like a tool.
Code words:
- Define the output format
- Define tone
- Define transformation type
- Reduce randomness
- Cut revision loops
Instead of rewriting long instructions every time, you apply a “mode.”
It feels like macros.
Pro Tips to Make Code Words Even More Powerful
Here’s where most people mess up.
1. Set Boundaries
Add instructions like:
- “Only use what I pasted.”
- “Do not assume missing details.”
- “If information is missing, say UNKNOWN.”
Otherwise, the model may fill gaps automatically.
Does ChatGPT Already Know These Code Words?
Short answer: No.
ChatGPT does not have built-in secret commands.
But it understands instructions extremely well.
When you clearly define what a code word means, it follows that instruction consistently.
The code word is just a trigger.
The clarity behind it is what matters.
