Best ChatGPT Prompt Templates for 2026: 25 Copy-Paste Prompts That Use Every Feature

I want to set up my Custom Instructions so ChatGPT gives me better responses by default. Help me fill out both sections.

About me:
- My role: [e.g., "product manager at a B2B SaaS startup"]
- My industry: [e.g., "fintech"]
- What I use ChatGPT for most: [e.g., "writing product specs, analyzing user data, drafting emails to stakeholders"]
- My communication style: [e.g., "direct, no fluff, data-driven"]
- Technical level: [e.g., "comfortable with SQL and basic Python, not a developer"]
- Tools I use daily: [e.g., "Notion, Figma, Linear, Slack, Google Sheets"]

Based on this, write me:
1. A "What would you like ChatGPT to know about you?" paragraph (under 1500 characters)
2. A "How would you like ChatGPT to respond?" paragraph (under 1500 characters)

Make the response instructions specific. Not "be concise" — something like "default to bullet points, lead with the recommendation, include tradeoffs only when I ask."

I want to teach you key facts about me and my work so you remember them across conversations. Store these:

Professional context:
- I work as [role] at [company]
- My team is [size] people, I report to [role]
- Current projects: [list 2-3 active projects]
- Tech stack: [tools and technologies]

Preferences:
- When I say "draft," I want a rough version I'll edit. When I say "write," I want something close to final.
- Default output format: [bullet points / paragraphs / tables]
- I prefer [metric system / imperial], [24h / 12h time], [DD/MM/YYYY / MM/DD/YYYY]
- My timezone: [timezone]
- When I share code, assume [language] unless I say otherwise

Working style:
- I value [brevity over completeness / thoroughness over speed]
- Don't hedge with "it depends" — give me your best recommendation, then mention caveats
- If I'm wrong about something, say so directly

Confirm what you've stored.

Help me create a Custom GPT for a specific workflow I repeat often.

Workflow I want to automate:
- What I do: [e.g., "review pull requests and write feedback comments"]
- How often: [e.g., "3-5 times per day"]
- Current process: [describe your manual steps]
- Inputs I'd provide: [e.g., "I paste in the diff or a link to the PR"]
- Output I want: [e.g., "structured review with severity levels, specific line references, and suggested fixes"]

Build me:
1. A name and description for the GPT
2. The full system prompt / instructions (be specific — include formatting rules, tone, what to include and what to skip)
3. Conversation starters (4 examples of how I'd kick off a session)
4. Recommended settings (web browsing on/off, code interpreter on/off, DALL-E on/off)

Make the instructions detailed enough that someone else on my team could use this GPT and get consistent results.

Write a blog post. But here's the thing — it needs to actually sound like a human wrote it.

Topic: [topic]
Target audience: [who reads this]
Goal: [educate / persuade / entertain / drive signups]
Word count: [target]
SEO keyword: [primary keyword]
My brand voice: [describe in 2-3 sentences, or paste your brand voice doc]
Something I personally know about this topic that most articles miss: [your unique angle or experience]

Writing rules:
- Start with a hook that's specific, not generic. No "In today's fast-paced world..." No "Have you ever wondered..."
- Vary sentence length. Some short. Some longer and more flowing with subclauses that give context.
- Use contractions (it's, don't, can't) — real people use them
- Include at least one analogy or metaphor that makes an abstract concept concrete
- Every section should have a clear "why should I care" before the "here's how"
- Use specific numbers and examples instead of vague claims
- End sections with a transition that makes the reader want to keep going
- No bullet-point lists longer than 5 items
- Write the meta description separately at the end (under 160 characters)

Write a social media thread for [platform: X/Twitter, LinkedIn, or Threads].

Topic: [what you want to talk about]
Hook angle: [hot take / personal story / surprising data / step-by-step / myth-busting]
Thread length: [5-10 posts]
CTA at the end: [follow / visit link / reply with their experience]
My voice: [how you sound on this platform — casual? authoritative? irreverent?]

Platform-specific rules:
- X/Twitter: Max 280 chars per post. Use line breaks for readability. First post = hook, last post = CTA. No hashtags in the thread, maybe 1-2 in a reply.
- LinkedIn: Longer format OK (up to 3,000 chars per post). Hook line must stop the scroll. Short paragraphs. Use → or • for lists, not generic bullets.
- Threads: Casual tone, 500 chars per post. More conversational than X. Emojis OK but don't overdo it.

The first post needs to make someone stop scrolling in under 2 seconds. If the hook doesn't punch, rewrite it until it does.

Write my weekly email newsletter.

Newsletter name: [name]
Audience: [subscribers — who are they and why'd they subscribe]
This week's main topic: [the big idea or theme]
Supporting items: [2-3 smaller things to include — links, quick tips, news, personal update]
Tone: [smart friend / industry insider / witty commentator / straight shooter]

Structure:
- Subject line (3 options: one curiosity, one benefit, one personal)
- Preview text (the snippet that shows in inbox — not just the first sentence repeated)
- Opening: Something that happened this week that connects to the main topic (personal observation, news, or reader question)
- Main section: The meat — insight, analysis, or actionable advice
- Quick hits: 2-3 smaller items with brief commentary
- Closer: One sentence that feels like a conversation, not a sign-off
- P.S. line: Optional but effective — use for CTA, fun fact, or teaser for next week

Keep total word count under [500 / 750 / 1000]. Every sentence should either inform, entertain, or both. Nothing that exists just to fill space.

I have a piece of long-form content that I need to repurpose into multiple shorter formats.

Original content:
"""
[Paste the full blog post, article, transcript, or video script — or attach the file]
"""

Create these assets from this content:
1. X/Twitter thread (7-10 posts) — extract the most tweetable insights
2. LinkedIn post — professional angle, 150-200 words, hook + insight + question
3. Instagram carousel script — 8-10 slides, each with a headline and 1-2 sentences
4. Email teaser — drive traffic back to the original piece, under 150 words
5. 3 standalone quote graphics — pull the most shareable one-liners, format as "quote — attribution"
6. YouTube Shorts / TikTok script — 45-60 seconds, one key takeaway, spoken delivery style

For each format, don't just shorten the original — rewrite it for how people consume content on that specific platform. A LinkedIn post reads different from a tweet reads different from a carousel.

Write sales copy for [product/service/offer].

What it is: [description]
Who it's for: [specific audience]
Price: [amount]
Main promise: [the transformation or outcome]
Biggest objection buyers have: [the thing that makes them hesitate]
Proof: [testimonials, stats, case studies, credentials]

Write:
1. Headline: benefit-driven, under 12 words, creates urgency without being sleazy
2. Subheadline: expand the promise, address the objection implicitly
3. Opening paragraph: describe the problem so vividly the reader thinks "that's exactly me"
4. Bridge: introduce the solution without naming it yet (build anticipation)
5. The offer: what they get, structured as a value stack
6. Objection handling: address the top 3 hesitations naturally within the copy (don't make a FAQ section — weave it in)
7. Social proof: testimonial placement with specific results
8. CTA: clear, specific, low-friction
9. Risk reversal: guarantee, free trial, or "worst case" framing
10. P.S.: one more reason to act now

This should read like a conversation with a trusted advisor, not a used car salesman.

I'm uploading a data file. Analyze it thoroughly and help me understand what's going on.

[Upload your CSV, Excel, or JSON file]

First, do a data health check:
- How many rows and columns?
- Any missing values? Where?
- Data types per column (are numbers stored as text? dates formatted consistently?)
- Any obvious outliers or errors?

Then analyze:
1. Summary statistics for all numerical columns
2. Distribution of categorical columns (top values, frequency)
3. Correlations between numerical columns — flag anything above 0.7 or below -0.7
4. Time-based trends (if there's a date column)
5. Top 3 most interesting patterns or anomalies you notice

Create 3-4 visualizations that tell the story of this data. Use clean, professional styling — no default matplotlib ugliness.

End with: "Based on this data, here are 3 things I'd investigate further and why."

Build a financial model based on these inputs.

[Upload spreadsheet or provide data below]

Business type: [e.g., "SaaS with monthly subscriptions"]
Revenue model: [how money comes in — subscriptions, one-time sales, usage-based, etc.]

Known inputs:
- Current MRR/revenue: [amount]
- Growth rate: [monthly/annual percentage]
- Customer acquisition cost (CAC): [amount]
- Lifetime value (LTV): [amount or "calculate it"]
- Monthly operating costs: [breakdown or total]
- Churn rate: [monthly percentage]
- [Add any other inputs you have]

Build:
1. 12-month and 36-month projection with monthly granularity
2. Key metrics dashboard: MRR, ARR, net revenue retention, LTV:CAC ratio, runway
3. Sensitivity analysis: what happens if churn increases 50%? If growth slows 30%?
4. Break-even analysis: when do we become profitable at current trajectory?
5. Charts: revenue over time, cost breakdown, unit economics

Generate the model as a downloadable Excel/CSV file with formulas I can edit.

I have survey results I need to make sense of.

[Upload CSV/Excel with survey responses]

Context:
- Survey purpose: [e.g., "customer satisfaction after product launch"]
- Number of respondents: [if not obvious from data]
- Key questions I care most about: [list 3-5]
- Any demographic segments to compare: [e.g., "new vs returning customers", "enterprise vs SMB"]

Analyze:
1. Response rate and completion rate
2. Quantitative results: averages, distributions, NPS/CSAT score if applicable
3. Cross-tabulation: how do answers differ between segments?
4. Open-ended response themes: group free-text answers into 5-7 themes with example quotes
5. Statistical significance: are the differences between groups real or just noise?
6. Visualizations: charts that I could put directly into a presentation

End with:
- Top 3 actionable insights (not just "satisfaction is 4.2/5" but "satisfaction drops to 3.1 for users who contact support more than twice")
- Recommended next steps based on the data

I'm uploading [number] documents that I need structured data from.

[Upload PDFs, images, or files]

For each document, extract:
- [Field 1, e.g., "candidate name"]
- [Field 2, e.g., "years of experience"]
- [Field 3, e.g., "key skills"]
- [Field 4, e.g., "education"]
- [Field 5, e.g., "most recent role and company"]
- [Add more fields as needed]

Output as:
- A clean table (one row per document)
- A downloadable CSV
- Flag any documents where extraction was uncertain

If any field is ambiguous or missing, put "UNCLEAR" rather than guessing.

I'm sharing a screenshot of [what it is — a Slack conversation, project board, email thread, dashboard, design mockup].

[Upload screenshot]

Based on what you see:
1. Summarize the key information in 3-5 bullet points
2. Identify any action items, deadlines, or decisions visible
3. Flag anything that looks problematic, incomplete, or needs attention
4. If there are any numbers or metrics visible, pull them into a clean table

Context that might help you interpret this:
- [e.g., "This is from our sprint planning board. We're in week 2 of a 3-week sprint."]
- [e.g., "The red items are overdue tasks."]

I'm sharing a screenshot of [a website, app screen, slide deck, marketing material, or design mockup].

[Upload image]

Review this design and give me specific, actionable feedback on:

1. Visual hierarchy: Is it clear where the eye goes first? Is the most important element the most prominent?
2. Readability: Font sizes, contrast, line spacing — can someone actually read this comfortably?
3. Layout: Is the spacing consistent? Does the grid feel intentional or random?
4. Color: Does the palette work? Any accessibility issues (contrast ratio problems)?
5. CTA clarity: If there's a call-to-action, is it obvious and compelling?
6. Mobile readiness: Would this work on a phone screen without major changes?

For each issue you find:
- Describe what's wrong
- Explain why it matters
- Give a specific fix

Be direct. "Looks good!" is useless. I need "the CTA button gets lost because it's the same visual weight as the navigation. Make it 2x larger and use your primary brand color."

I'm uploading a photo of [handwritten notes, a whiteboard, a printed document, a receipt, a business card].

[Upload image]

1. Transcribe ALL text you can read, maintaining the original structure (lists stay as lists, headers stay as headers)
2. Where handwriting is unclear, give your best guess in [brackets with ?], like [inventory?]
3. If this is a whiteboard or diagram, also describe the visual structure (arrows, boxes, connections)
4. Organize the transcription into clean markdown

After transcribing:
- [Optional: summarize the key points]
- [Optional: convert handwritten math into properly formatted equations]
- [Optional: restructure messy notes into a clean outline]

Research [competitor name] and give me a competitive intelligence briefing.

My company: [your company and what you do]
The competitor: [competitor name and URL]
What I need to understand: [pricing strategy / product features / content strategy / hiring patterns / market positioning]

Research:
1. Current product offerings and pricing (check their website, not cached data)
2. Recent news, announcements, or launches (last 90 days)
3. How they position themselves vs competitors (read their homepage, about page, and recent blog posts)
4. Their apparent target customer (based on language, case studies, and testimonials visible)
5. Job postings — what roles are they hiring for? (This reveals strategic priorities)
6. Social media presence — where are they active, what's their engagement like?

Compile into:
- Executive summary (5 sentences)
- Strengths and weaknesses (3 each, with evidence)
- Opportunities for us (where they're weak or absent)
- Threats (where they're strong or gaining ground)
- Recommended actions (3 specific things we should do in response)

Cite your sources with URLs so I can verify.

I need you to research a topic and synthesize what you find into something I can actually use.

Topic: [what you need researched]
Depth: [surface-level overview / working knowledge / deep expertise]
Use case: [why I need this — making a decision, writing a report, building a strategy, personal learning]
Time sensitivity: [does this need to be current 2026 data, or is evergreen knowledge fine?]

Research approach:
1. Find 5-8 high-quality sources (prioritize: official documentation, reputable publications, academic sources, industry reports)
2. Identify points of consensus — what do most sources agree on?
3. Identify points of disagreement — where do experts differ, and why?
4. Note any data points or statistics worth citing

Deliverable:
- Key findings (organized by subtopic, not by source)
- Data table if applicable
- "What this means for [my specific use case]" section
- Limitations — what couldn't you find, or what's uncertain
- Source list with URLs

I need to verify the accuracy of some claims before I publish or share them.

Claims to verify:
1. "[Claim 1]"
2. "[Claim 2]"
3. "[Claim 3]"
[Add more as needed]

For each claim:
- VERDICT: True / Mostly True / Misleading / False / Unverifiable
- EVIDENCE: What sources say, with links
- CONTEXT: Is the claim technically true but misleading? Is it outdated? Is it true in some contexts but not others?
- CORRECTION: If false or misleading, what's the accurate version?

Source requirements:
- At least 2 independent sources per claim
- Prefer primary sources (studies, official data) over secondary coverage
- If a claim references a specific study or stat, find the original source — not just someone else citing it
- Flag anything where you can only find one source or where sources conflict

I'd rather you say "I can't verify this" than make something up.

I have a bug I can't figure out. Help me find and fix it.

Language: [language]
Framework/environment: [e.g., "React 18 with TypeScript", "Python 3.11 with FastAPI", "Node.js 20 with Express"]

The code:
```[language]
[paste your code]

For a full-featured code review workflow, check out our [Claude vs ChatGPT for Coding comparison](/blog/claude-vs-chatgpt-coding/) to see which AI handles different coding tasks better.

### 20. Build a Feature from Scratch

### 21. API Integration Helper

### 22. Code Review

[paste your code]

## Creative and Image Generation

DALL-E is built into ChatGPT. These templates get you better images than "draw me a cat."

### 23. Product Mockup Generator

### 24. Social Media Visual Concept

## Deep Reasoning with o3 / o4-mini

ChatGPT's reasoning models (o3, o4-mini) actually think through problems step by step. They're slower but dramatically better for complex tasks. Use these when the answer isn't obvious.

### 25. Complex Decision Framework

## How to Use These Templates

**Step 1: Set up Custom Instructions first.**
Templates 1 and 2 take five minutes and make everything else work better. Do them before anything else.

**Step 2: Match the template to ChatGPT's features.**
Uploading a file? Use the Code Interpreter templates (9-12). Sharing a screenshot? Use the Vision templates (13-15). Need current information? Use the Browsing templates (16-18). For complex reasoning, switch to the o3 or o4-mini model and use template 25.

**Step 3: Fill in every bracket.**
The placeholders exist for a reason. "[my product]" gives you generic output. "[CloudSync, a file backup tool for remote design teams that auto-syncs Figma, Sketch, and Adobe files]" gives you output you can actually use.

**Step 4: Iterate.**
First output is a draft. Tell ChatGPT "tighten the opening," "this reads too formal," or "cut 30% and keep only the strongest points." The best results come from conversation, not one-shot prompts.

**Step 5: Try the same prompt in different models.**
GPT-4o is fastest for everyday tasks. o3 is better when you need careful reasoning. o4-mini is a good middle ground. Experiment — the model choice often matters as much as the prompt.

For more on getting better AI output in general, our [prompt engineering guide](/blog/prompt-engineering-beginners-guide/) covers the foundational techniques. And if you're curious how ChatGPT stacks up against the alternatives, our [ChatGPT vs Claude vs Gemini comparison](/blog/chatgpt-vs-claude-vs-gemini/) breaks it down by use case.

## Get More Templates

These 25 templates cover ChatGPT's core feature set in 2026 -- from custom instructions to reasoning models. But ChatGPT can do a lot, and you might need something more specific.

Browse our full library for more:
- [Blog Post Writer](/skills/ai-creative/blog-post-writer/) -- full-featured blog content creation
- [Landing Page Copywriter](/skills/ai-creative/landing-page-copywriter/) -- conversion-focused sales pages
- [Content Research Writer](/skills/ai-creative/content-research-writer/) -- research-backed long-form content
- [Professional Email Writer](/skills/productivity/professional-email-writer/) -- polished email drafting
- [Cold Email Outreach Pro](/skills/productivity/cold-email-outreach/) -- sales prospecting sequences
- [Full skill library](/skills/) -- 1,000+ skills across every category

Every skill is a copy-paste prompt. No accounts, no subscriptions. Just find what you need and use it.

ChatGPT gets more capable with every update. Your prompts should too. These templates are the starting point -- take them, customize them, and build workflows that actually save you time.