Prompt Library

Blog Post Outline Generator

Planning long-form content before writing

Act as a content strategist. Create a blog post outline for the topic "[TOPIC]" targeting [AUDIENCE]. Include: a hook-driven title…

Tone Rewriter

Adjusting tone without rewriting from scratch

Rewrite the following text to sound more [tone: e.g. confident / friendly / formal], keeping the same meaning and roughly the same…

Cold Email Opener Generator

Outbound sales / partnership outreach

Write 5 different cold email opening lines for reaching out to [ROLE] at [TYPE OF COMPANY] about [YOUR OFFER]. Each opener should …

Competitor Positioning Snapshot

Competitive research with citations. Use this before a positioning workshop, a pricing review, or writing comparison copy — the table format makes it easy to spot where a rival is genuinely stronger versus where they are just louder. Always run it on a search-grounded model like Perplexity so every claim carries a source and a date you can verify.

Research [COMPETITOR NAME]'s current positioning, pricing tiers, and most recent product announcements (last 6 months). Summarize …

ClaudeProductivity

Meeting Notes to Action Items

Turning messy notes into a shareable recap

Convert these raw meeting notes into: (1) a 3-sentence summary, (2) a bulleted list of decisions made, (3) a bulleted list of acti…

ChatGPTCoding

SQL Query Explainer

Understanding or debugging unfamiliar queries. Ideal when you inherit a legacy codebase, review a teammate's pull request, or hit a slow query you did not write. Asking for a plain-English walkthrough before the optimisation notes means you understand the intent first, so you can tell a genuine bug from a deliberate design decision.

Explain what this SQL query does in plain English, step by step, then flag any potential performance issues or bugs. [PASTE SQL]…

ClaudeCoding

Code Review Pass

Pre-review pass before submitting a pull request

Review this code as a senior engineer doing a PR review. Flag: (1) bugs or edge cases, (2) readability issues, (3) any security co…

n8n + ChatGPTAutomation

Weekly Automation Digest

Automated weekly team update via n8n + Slack

Summarize this week's data into a short digest for the team. Input: {{$json.metrics}}. Output format: 3 bullet highlights, 1 conce…

n8n + ClaudeAutomation

RSS-to-Social Repurposer

Auto-posting new articles to social channels

New RSS item: {{$json.title}} — {{$json.content}}. Turn this into: (1) one tweet under 280 characters with a link placeholder, (2)…

GrokMarketing

Ad Angle Brainstorm

Generating creative directions before writing ad copy

Give me 8 different marketing angles for [PRODUCT] — not copy yet, just the underlying angle/hook idea for each (e.g. "fear of mis…

Hugging Face (Stable Diffusion)Image Generation

Product Photo Prompt (Studio Style)

Clean product shots for a store listing. Useful when you need placeholder or concept imagery before a real photoshoot, or for a product that does not physically exist yet. Keep the subject description first and the styling details after — most image models weight earlier tokens more heavily, so leading with the object keeps it recognisable.

Professional studio product photography of [PRODUCT], centered on a seamless [COLOR] background, soft box lighting from top left, …

Hugging Face (Stable Diffusion)Image Generation

Brand Mood Board Prompt

Visual direction before a brand shoot. Generating a mood board first gives designers and photographers something concrete to react to, which is far faster than describing a feeling in a brief. Run it several times with different adjectives and keep the two or three frames that feel closest — the misses are as informative as the hits.

[STYLE, e.g. minimalist Scandinavian] mood board for a [INDUSTRY] brand, muted color palette of [COLORS], flat lay composition, na…

YouTube Video Idea Generator

Content calendar brainstorming

Generate 10 YouTube video ideas for a channel about [NICHE], targeting [AUDIENCE]. For each: a title under 60 characters and a one…

Fact-Check This Claim

Verifying a statistic or claim before publishing. Run this on anything you did not source yourself, especially numbers pulled from a chat model — statistics are exactly the kind of detail models hallucinate confidently. Use a search-grounded tool so you get a live primary source and a date, and treat "no source found" as a reason to cut the claim rather than to keep looking.

Fact-check this claim: "[CLAIM]". Tell me if it's accurate, partially accurate, or false, and cite the sources you used with publi…

ClaudeProductivity

Daily Priority Planner

Quick daily task triage

Here is my task list for today: [PASTE TASKS]. Sort into: Must do today, Should do today, Can wait. Base it on impact and urgency,…

The 5-Part Prompt Framework — ChatGPT

Every strong prompt combines five parts: Role (who the AI should act as), Context (background info it needs), Task (the specific ask), Format (how you want the output structured), and Constraints (length, tone, things to avoid).

You are a senior email marketer. Context: I run a 3-person SaaS startup selling project management software to freelancers. Task: …

The 5-Part Prompt Framework — Claude

Role: You are a meticulous technical editor. Context: The attached text is a beginner tutorial on Python loops. Task: Rewrite it f…

Zero-Shot vs Few-Shot Prompting — ChatGPT

Zero-shot means asking directly with no examples.

Classify each review as Positive, Negative, or Neutral. Review: "Delivery was fast but the box arrived damaged." -> Neutral Revie…

GrokFundamentals

Zero-Shot vs Few-Shot Prompting — Grok

Zero-shot means asking directly with no examples.

Here are 3 examples of turning a boring headline into a punchy one: "5 Tips for Productivity" -> "Stop Wasting Your Mornings: 5 Pr…

Chain-of-Thought Prompting — ChatGPT

Asking a model to "think step by step" before answering measurably improves accuracy on reasoning, math, and multi-step logic tasks because it forces the model to externalize intermediate steps instead of jumping to a guess..

A customer has a $1,200 budget. Product A costs $340 and lasts 2 years. Product B costs $610 and lasts 4 years. Which has the bett…

Chain-of-Thought Prompting — Claude

Before answering, list the 3 possible root causes of this bug, evaluate each against the stack trace below, then state which is mo…

Structured Output (JSON & Tables) — ChatGPT

When you need output your app or spreadsheet can actually use, explicitly request a schema.

Extract the following fields from this job listing as JSON only, no extra text: { "title": string, "company": string, "location": …

Hugging Face (open-source)Fundamentals

Structured Output (JSON & Tables) — Hugging Face (open-source)

When you need output your app or spreadsheet can actually use, explicitly request a schema.

System: Respond only with valid JSON matching this schema: {"summary": string, "sentiment": "positive"|"neutral"|"negative", "keyw…

PerplexityFundamentals

ChatGPT vs Claude vs Perplexity vs Grok — Perplexity

The same task often needs a slightly different prompt style per model.

What are the three most significant changes to [topic] in the last 6 months? Cite sources for each claim and note the publish date…

GrokFundamentals

ChatGPT vs Claude vs Perplexity vs Grok — Grok

The same task often needs a slightly different prompt style per model.

Give me a quick, no-fluff rundown of what people are currently saying about [topic] — what's the general sentiment right now?…

Role + Constraints to Kill Vagueness — ChatGPT

When answers are generic, the fix is almost always more specificity, not a longer prompt.

You are a conversion copywriter for B2B SaaS. Rewrite this headline for a CFO audience. Constraints: under 12 words, lead with a q…

Role + Constraints to Kill Vagueness — Claude

When answers are generic, the fix is almost always more specificity, not a longer prompt.

Act as a plain-language editor for a 10-year-old reader. Rewrite the paragraph below at a 5th-grade reading level, keep it under 8…

Ask-Before-Answer for Ambiguous Tasks

For open-ended requests, instruct the model to ask clarifying questions first.

I want help planning a product launch. Before you give any plan, ask me up to 5 clarifying questions about my product, audience, t…

Self-Grading With a Rubric

Have the model score its own draft against an explicit rubric, then revise to hit the top score.

Write a 100-word product pitch for [PRODUCT]. Then score it 1-5 on: clarity, specificity, and persuasiveness. If any score is belo…

Any modelFundamentals

Prompt Chaining Architectures

Experts rarely use one mega-prompt.

STEP 1 (extract): From the customer interview below, extract pains, desired outcomes, and objections as JSON: {"pains":[],"outcome…

Meta-Prompting: Prompts That Write Prompts

The fastest way to a great prompt is asking the model to write and critique one.

You are a prompt engineer. Task: I need a reusable prompt that turns raw meeting notes into a client-ready status email. Draft the…

Any modelFundamentals

Evaluating Prompts Like an Engineer

A prompt is good when it performs consistently, not when one output impressed you.

Here are 2 versions of a prompt and 5 test inputs. Run each version mentally against each input. Build a table: input × version → …

Any modelFundamentals

Project: A 3-Step Prompt Chain for Your Real Work

Deliverable: a documented, tested 3-step prompt chain that automates one recurring task from your actual work or studies. Steps: 1.

Act as my project reviewer. Here is my 3-step prompt chain and one test input with its output. Grade against: single-responsibilit…

ChatGPTAutomation

Critique-and-Revise Loops — ChatGPT

A loop is simply feeding a model's own output back to it (or to a second prompt) for review and improvement, repeated until a quality bar is met.

Step 1 — Draft: Write a 150-word product description for [product]. Step 2 — Critique: Now critique your own draft as a skeptical …

ClaudeAutomation

Critique-and-Revise Loops — Claude

A loop is simply feeding a model's own output back to it (or to a second prompt) for review and improvement, repeated until a quality bar is met.

Write the function requested below. Then, in a second pass, review your own code for edge cases and bugs as if you were a strict s…

ChatGPT / Agent toolsAutomation

The ReAct Pattern (Reason + Act)

ReAct loops interleave reasoning ("what should I do next?") with actions (calling a tool, searching, running code), then feed the result back in before reasoning again.

You have access to a "search" tool and a "calculator" tool. For the question below, reason about which tool (if any) you need, sta…

ClaudeAutomation

Giving Loops Memory

For loops that run over many steps (research agents, multi-turn assistants), summarize prior steps into a short running memory instead of replaying the full history every time — this keeps context small and cheap..

Here is a running summary of research so far: [summary]. Here is the newest finding: [new info]. Update the summary to under 150 w…

Planner–Worker–Reviewer Teams

The most robust agent architecture is also the simplest: a Planner decomposes the goal into tasks, Workers execute one task each with narrow context, and a Reviewer checks results against acceptance criteria before anything ships.

PLANNER: Decompose this goal into 3-5 independent tasks with acceptance criteria each. Goal: [GOAL]. WORKER (per task): Complete e…

Loop Budgets and Failure Handling

Production loops need budgets (max iterations, max cost, max time) and explicit failure paths: what happens when the Reviewer fails a result three times? Options: escalate to a human, fall back to a simpler method, or return a partial result flagged as unverified.

Design the failure policy for this loop: [DESCRIBE LOOP]. Specify: max retries per step, what triggers human escalation, what a sa…

Project: Draft–Critique–Revise Content Machine

Deliverable: a working 3-role loop (writer, critic, editor) that produces a publishable piece of content on any topic you give it. Steps: 1.

CRITIC role for my loop: Score this draft 1-5 on: hook strength, specificity, structure, and CTA clarity. For any score under 4, g…

n8n / HTTP Request nodeAutomation

Connecting an LLM to n8n

n8n connects to AI models either through dedicated nodes (OpenAI, Anthropic community nodes) or a generic HTTP Request node calling any model's API.

Method: POST. Body (JSON): { "model": "your-model", "messages": [{"role":"system","content":"You summarize support tickets in 1 se…

ChatGPT (inside n8n)Automation

Auto Content Repurposing Pipeline

Trigger: new blog post published (RSS or webhook) -> AI node: generate a Twitter/X thread, LinkedIn post, and email blurb from the article -> Route each to the right channel.

Turn this blog post into 3 outputs: (1) a 5-tweet thread, (2) a 150-word LinkedIn post, (3) a 2-sentence email teaser with a link …

Claude (inside n8n)Automation

AI Email Triage & Auto-Responder

Trigger: new email in inbox -> AI node classifies intent (support, sales, spam, urgent) -> conditional branches route to auto-reply, Slack alert, or human queue..

Classify this email into exactly one category: "support", "sales_lead", "spam", "urgent_escalation". Then draft a 2-sentence ackno…

ChatGPT (inside n8n)Automation

Error Handling & Idempotency in n8n

Real automations fail: APIs time out, models return malformed JSON, webhooks fire twice.

Return ONLY valid JSON: {"category": string, "reply": string, "confidence": number 0-1}. If you cannot classify confidently, set c…

Claude (inside n8n)Automation

Human-in-the-Loop Approval Gates

The highest-trust automations pause for approval at the moments that matter: before sending externally, before spending money, before deleting anything.

Draft the approval message for this automated action. Include: what will happen, to whom, the exact content being sent, and two op…

Project: Your First Production n8n Workflow

Deliverable: one n8n workflow, running on a schedule or trigger, that saves you real time every week. Steps: 1.

Here is my n8n workflow description and its AI prompt. Review it as an automation engineer: where will it break first, what input …

PerplexityMarketing

Finding Your Audience's Real Language

Before writing copy, use AI to mine the language your audience actually uses — their complaints, objections, and desires — so your copy sounds like them, not like a brochure..

Find recent (last 12 months) online discussions, reviews, and forum posts where people talk about problems with [your product cate…

AIDA and PAS Prompts — ChatGPT

AIDA (Attention, Interest, Desire, Action) and PAS (Problem, Agitate, Solution) are proven copy structures.

Write a landing page hero section using the PAS framework for [product]. Problem: [target pain point]. Agitate: make the cost of i…

AIDA and PAS Prompts — Claude

AIDA (Attention, Interest, Desire, Action) and PAS (Problem, Agitate, Solution) are proven copy structures.

Using AIDA, write a 4-part cold outbound email for [product] targeting [ICP]. Keep each part to 1-2 sentences. End with a low-fric…

AI-Assisted Audience Segments

Feed AI a description of your customer base or CRM export summary and ask it to propose segments with distinct messaging angles — a fast way to go from "everyone" to targeted campaigns..

Here is a summary of our customer base: [paste summary/stats]. Propose 3 distinct audience segments. For each, give: segment name,…

AI-Assisted Positioning Statements

Positioning comes before copy: who it's for, what category you're in, what alternative you replace, and the one difference that matters.

Using April Dunford's positioning components, generate 3 distinct positioning options for [PRODUCT]. For each: target segment, mar…

Full-Funnel Message Mapping

Different funnel stages need different messages: problem-aware content at top, comparison and proof in the middle, urgency and risk-reversal at the bottom.

Build a funnel message map for [PRODUCT] targeting [ICP]. For each stage (unaware, problem-aware, solution-aware, deciding): the c…

Project: Research → Position → Launch Assets

Deliverable: a complete mini campaign kit for a real or invented product. Steps: 1.

Here is my positioning statement and campaign assets. Check consistency: does every asset express this positioning? Flag any asset…

Generating High-Variance Ad Sets — ChatGPT

Ad platforms reward variety for testing.

Write 6 Google Ads headlines (max 30 characters each) for [product/offer]. Use a different angle for each: (1) price/value, (2) ur…

GrokMarketing

Generating High-Variance Ad Sets — Grok

Ad platforms reward variety for testing.

Give me 5 scroll-stopping Meta ad primary text openers (first line only, this is what shows before "See more") for [product]. Make…

Reading Ad Performance Data with AI

Paste raw campaign metrics and ask AI to find the story in the numbers — which angle, audience, or creative is actually winning, and what to do next..

Here is last week's ad performance data (CTR, CPC, CPA, spend by ad set): [paste table]. Identify: (1) the best and worst performi…

Creative Testing Systems (Not One-Off Tests)

Scaling accounts run a creative system: a backlog of angles ranked by hypothesis strength, batches of 3-5 new creatives weekly, and a kill/scale rule applied without emotion (e.g., kill under 0.8× account-average CTR after 2k impressions; scale winners into new formats).

Maintain my creative backlog. Here are past results: [PASTE angle → CTR/CPA table]. Rank remaining untested angles by likely perfo…

Landing Page ↔ Ad Message Match

Half of "ad fatigue" is actually message mismatch: the ad promises one thing, the landing page opens with another, and CVR dies.

Ad: [PASTE AD COPY]. Landing page above-the-fold: [PASTE]. Score message match 1-5 on: promise continuity, audience continuity, an…

Project: Design, Launch (or Simulate) and Read a Test

Deliverable: a complete documented ad test — real if you have an account, fully simulated if not. Steps: 1.

Role-play 20 people matching this profile: [ICP]. Show each of these 4 ads. For each person output: which ad stopped them, and the…

Killing the "AI Voice"

Generic AI writing tends to over-use certain patterns (excessive "moreover," perfectly balanced sentences, hedge-everything tone).

Here is a sample of my writing style: [paste 2-3 paragraphs you wrote]. Write a new [blog post/caption] about [topic] matching thi…

Hugging Face (Stable Diffusion)Writing

Structuring Image Prompts

Good image prompts (for DALL·E, Stable Diffusion / Hugging Face models, Midjourney-style tools) generally follow: Subject -> Setting -> Style/medium -> Lighting/mood -> Composition/technical details..

A minimalist product photo of a matte black wireless earbuds case, floating on a soft gradient blue background, studio lighting, s…

One Article, Five Formats

Turn a single long-form piece into a newsletter blurb, Twitter/X thread, LinkedIn post, Instagram caption, and short-form video script — with one well-structured prompt..

Repurpose the article below into: (1) a 60-word newsletter blurb, (2) a 6-tweet thread, (3) a LinkedIn post under 150 words, (4) a…

Building a Reusable Brand Voice Prompt

Instead of re-describing your tone every time, create a single "voice block" you paste at the top of every content prompt.

Create a reusable "brand voice" instruction block for my brand based on these samples: [PASTE 2-3 SAMPLES]. Output it as a single …

Batch-Creating a Week of Content

Once you have a voice block and a topic, you can generate a whole content calendar in one structured request — then refine individual pieces..

[PASTE VOICE BLOCK] Using that voice, create a 5-day content plan on [TOPIC]. For each day give: platform, hook, main point, and …

The Two-Pass Edit: Structure Then Line — Claude

Professional editing is two separate passes: a structural pass (does the argument build? is anything missing or redundant?) and a line pass (rhythm, word choice, cuts).

PASS 1 — structural edit only. Do not touch sentences. Output: the piece's argument in one line, a list of sections that don't ser…

The Two-Pass Edit: Structure Then Line — Claude (2)

Professional editing is two separate passes: a structural pass (does the argument build? is anything missing or redundant?) and a line pass (rhythm, word choice, cuts).

PASS 2 — line edit only, structure is final. Tighten every sentence, cut 15% of words, kill clichés, vary rhythm. Do not add ideas…

Any modelWriting

Original Insight: The Thing AI Can't Fake

AI averages the internet; audiences reward what the average doesn't contain — your data, your contrarian take, your first-hand story.

Interview me to extract original material for a piece on [TOPIC]. Ask me 5 questions, one at a time, digging for: a first-hand exp…

Project: One Pillar Piece + Four Repurposed Assets

Deliverable: a published (or publish-ready) pillar piece with its full repurposing kit, all in your voice. Steps: 1.

Here are my pillar piece and 4 repurposed assets. Check: (1) does each keep the core insight intact, (2) does each respect its for…

Hook-First Script Structure

The first 5-10 seconds decide retention.

Write a YouTube video script outline on [topic] for a [X]-minute video. Structure: (1) 3 hook options for the first 10 seconds, (2…

Title & Thumbnail Text Pairing — Claude

Titles and thumbnail text should never repeat the same words — they should work together to create curiosity.

For a video about [topic], generate 5 title + thumbnail-text pairs. Rule: the title and thumbnail text must never repeat the same …

PerplexityWriting

Title & Thumbnail Text Pairing — Perplexity

Titles and thumbnail text should never repeat the same words — they should work together to create curiosity.

What are people currently searching for related to [topic] on YouTube? List related search terms and any recent (last 3 months) sp…

Reading Your Analytics with AI

Paste your YouTube Studio export and ask AI to find patterns across your best and worst performing videos — topic, length, hook style, and thumbnail pattern..

Here is a table of my last 15 videos with views, average view duration %, and CTR: [paste data]. Identify patterns: what do my top…

Scripting for the Retention Graph

Retention drops happen at predictable points: the 30-second mark (promise not yet paid), mid-video (a segment overstays), and any moment of throat-clearing ("before we start…").

Restructure this script for retention. Rules: pay the core promise a first time within 60 seconds, open a loop before each section…

Series & Format Design

Channels grow on repeatable formats, not one-off videos: a recognizable structure (same open, same segments, same visual grammar) lowers production cost and trains the audience to return.

Design a repeatable format spec for a series on [TOPIC/NICHE]: name, 15-second cold-open pattern, 3 fixed segments with target len…

Project: One Video, Fully Engineered

Deliverable: a complete production package for one real video on your (or an invented) channel. Steps: 1.

Generate the metadata kit for this video. Title: [TITLE]. Script summary: [SUMMARY]. Output: a 150-word description with keywords …

ChatGPTResearch

Mining Keywords and Search Intent — ChatGPT

AI is excellent at expanding a seed topic into keyword clusters and classifying search intent (informational, commercial, transactional, navigational).

I run a site about [topic]. Generate 5 keyword clusters around it. For each cluster give: a pillar keyword, 5 supporting long-tail…

Mining Keywords and Search Intent — Perplexity

AI is excellent at expanding a seed topic into keyword clusters and classifying search intent (informational, commercial, transactional, navigational).

What questions are people currently asking online about [topic]? List 15 real questions from forums, Reddit, and Q&A sites from th…

ClaudeResearch

Competitor Content Gap Analysis

Paste a competitor's page structure (headings, topics covered) alongside yours and ask AI to find gaps — subtopics they cover that you don't, questions they answer, and angles they missed that you can own..

Here are the H2/H3 headings of the top 3 ranking pages for "[keyword]": [paste headings]. Here is my page outline: [paste outline]…

ChatGPTResearch

Titles, Metas & Headings That Rank

Generate title tag and meta description variants that include the target keyword naturally, fit character limits, and give a reason to click.

Write 5 title tags (max 60 characters) and 3 meta descriptions (max 155 characters) for a page targeting "[keyword]". Rules: inclu…

Getting Cited by ChatGPT & Perplexity — Perplexity

AI answer engines favor content that is clearly structured, directly answers questions, cites data, and demonstrates first-hand expertise.

Ask: "[a question your content answers]". Note which sources get cited in the answer. Then analyze: what do those cited pages have…

ClaudeResearch

Getting Cited by ChatGPT & Perplexity — Claude

AI answer engines favor content that is clearly structured, directly answers questions, cites data, and demonstrates first-hand expertise.

Rewrite the opening of this article so the first 2 sentences directly and completely answer the question "[target question]" — the…

ClaudeCoding

What RAG Actually Does

RAG (Retrieval-Augmented Generation) retrieves the most relevant passages from your documents and injects them into the prompt before the model answers.

Answer the question below using ONLY the provided context. If the context doesn't contain the answer, say "Not found in the provid…

ChatGPTCoding

Chunking: Why Split Size Matters

Documents are split into chunks before embedding.

Split this document into retrieval-ready chunks. Rules: each chunk 200-400 words, never split mid-paragraph, prepend each chunk wi…

ChatGPT / ClaudeCoding

Custom GPTs & Claude Projects

Both ChatGPT (Custom GPTs) and Claude (Projects) let you upload documents and set standing instructions — a no-code RAG setup.

You are the internal support assistant for [company]. Answer questions using the uploaded documents as your only source of truth. …

ClaudeCoding

Writing Docs That Retrieve Well

Retrieval quality starts with document quality.

Restructure this document for AI retrieval: break it into sections with descriptive question-style headings, one topic per section…