# Gemini Prompt Library These are frameworks and meant to be modular, especially the build prompts vs. research prompts. Add-on modules that you see fit and don't forget to test and iterate. Remember, the best and most consistent output is a by product of the input; which is prompting + context. {% hint style="info" %} **Become a 10x Content Engineer with Cuppa AI:** [**Learn More HERE**](https://cuppa.ai) **(FREE TRIAL & WORKFLOW SESSION)** {% endhint %} Global Instruction (The "System" Prompt)\ \&#xNAN;*For the best results, paste this block at the very start of your chat session or your extra prompt in presets with Cuppa. This configures the AI's "brain" with your specific rules before you ask it to do any specific tasks.* ``` **SYSTEM CONFIGURATION: 2026 SEO & SEMANTIC ARCHITECT** **Role:** You are an elite SEO Strategist and Content Editor. You do not write generic content; you write high-performance, NLP-optimized assets. **CORE WRITING PRINCIPLES (Strict Adherence):** 1. **Sentence Structure:** Write in clear, direct sentences using **Subject–Verb–Object** order. Avoid complex clauses that confuse NLP parsers. 2. **Voice:** Conversational but authoritative. Write like you're explaining insights to a smart friend. Use contractions, rhetorical asides, and occasional humor. 3. **The "Anti-Fluff" Filter:** Ensure every sentence either teaches, makes the reader feel something, or entertains. 4. **BANNED WORDS (Immediate Fail Condition):** Do NOT use: *meticulous, navigating, complexities, realm, bespoke, tailored, ever-evolving, in the world of, unlock the secrets, unveil, robust, game-changer, tapestry.* 5. **Audience:** {Industry} audience in {GEO}. Reading Level: Grade {ENTER GRADE}. **SEO & NLP LOGIC:** 1. **Keyword Usage:** * Main keyword MUST appear in the **first 25 words**. * Keyword Density: **1-2%** (10-20 times per 1,000 words). * Integrate semantic variants naturally. 2. **Semantic Triples:** Use the structure of "Entity A -> Affects -> Entity B" logic in your prose. Do not list triples; weave them into cause-and-effect sentences. 3. **The 60/40 Rule:** 60% Machine Logic (clear, unambiguous syntax) + 40% Human Engagement (relatability). **Acknowledge this configuration and await my first prompt.** ``` *** #### 1. Foundational SEO Prompts (Strategy & Intent) *These prompts establish the semantic map and user journey.*\ **1.1 The "Micro-Intent" & Journey Mapper**\ **Goal**: Move beyond "Informational" to specific psychological states.
``` **Task:** Analyze the following keywords. Map them to the specific stage of the Customer Journey (Awareness, Consideration, Decision, Retention). **Analysis Framework:** 1. **Primary Keyword:** The keyword provided. 2. **Micro-Intent:** Go deeper than "Informational." Is it "Troubleshooting," "Comparison," "Definition," or "Purchase Ready"? 3. **The "Job to be Done":** What specific problem is the user trying to solve immediately? 4. **Semantic Triples:** Identify the core relationship needed to answer this query (e.g., "[Product] solves [Pain Point]"). 5. **Content Format:** (e.g., Listicles, How-to, Calculator, Product Page). **Keywords:** [PASTE KEYWORDS] **Output:** Present this as a Markdown Table. ``` **1.2 Semantic Keyword Cluster Builder**\ **Goal**: Build authority using Hub-and-Spoke logic and Semantic Triples.
``` **Task:** Create a keyword cluster around the topic: [TOPIC]. **Requirement:** Organize this cluster using **Semantic Triples logic**. I do not just want keywords; I want to know the *relationship* between the sub-topic and the main topic. **Output Format:** 1. **Cluster Name (The Hub):** [TOPIC] 2. **Sub-Topic (The Spoke):** [KEYWORD] 3. **The Relationship (The Triple):** Describe how the Spoke relates to the Hub using Subject -> Predicate -> Object logic (e.g., "Content Marketing [Subject] builds [Predicate] Topical Authority [Object]"). 4. **Primary Intent:** (Informational/Commercial) 5. **Projected Traffic Value:** (High/Med/Low) **Topic:** [PASTE TOPIC] ``` *** #### 2. On-Page Optimization Prompts *Optimizing titles, headers, and structure for NLP clarity.*\ **2.1 The "Hook & Click" Title Generator**\ **Goal**: High CTR + Keyword placement.
``` **Task:** Generate 10 SEO Title Tags for [KEYWORD]. **Constraints:** - **Front-Load:** Main keyword must appear in the first 50% of the title. - **Psychological Trigger:** Use Curiosity, Negativity Bias, or Specificity (Numbers). - **Length:** Under 60 characters (pixel width optimized). - **Differentiation:** Avoid generic "Ultimate Guide" titles unless necessary. - **Banned Words:** Do not use "Unlock" or "Secrets." **Output:** 10 Options with a brief explanation of the trigger used. ``` **2.2 The Semantic Header (H2/H3) Builder**\ **Goal**: Ensure headers imply relationships (Triples) rather than just labeling sections.\ \*Check out Semantic Triples Prompting section as well. Highly recommend for body sections ``` **Task:** Create a clean outline for [KEYWORD]. **H2/H3 Rules:** 1. **Triples Logic:** Write headers that imply a complete thought or relationship. * *Bad:* "Benefits" * *Good:* "How [Keyword] Increases Revenue" (Subject -> Predicate -> Object) 2. **Keyword Integration:** Use the main keyword or semantic variants in 50% of the headers. 3. **Zero Fluff:** No "Introduction" or "Conclusion" headers. Use descriptive headers instead. **Output:** A structured H1, H2, and H3 outline. ``` *** 3\. Content Production Prompts (The "Core" Engine)\ \&#xNAN;*These are the most critical prompts. They include your specific constraints regarding density, placement, and tone.*\ 3.1 The "NLP-Ready" Article Draft\ Goal: A full draft that passes your 8/10 quality score. ``` **Task:** Write a complete SEO article for the keyword: [KEYWORD]. **CRITICAL EXECUTION RULES:** 1. **First 25 Words:** You MUST include [KEYWORD] in the first 25 words of the intro. 2. **Keyword Density:** Maintain a 1-2% density. (For a 1000-word article, use the keyword 10-20 times). Do not stuff it; weave it naturally. 3. **Sentence Structure:** Use Subject-Verb-Object order. Keep sentences precise. 4. **Banned Words Check:** Do NOT use: *meticulous, navigating, complexities, realm, bespoke, tailored, ever-evolving, unlock, robust.* 5. **Semantic Triples:** Ensure the body content explains relationships clearly (e.g., "[Entity A] enables [Entity B]"). **Tone & Style:** - **Voice:** Conversational, authoritative, "smart friend." - **Formatting:** Short paragraphs (max 3 lines). Bullet points for data. - **Reading Level:** Grade {INSERT GRADE}. **Outline to Follow:** [PASTE OUTLINE] ``` 3.2 The "Anti-AI" Content Rewriter\ Goal: Fix existing content that sounds robotic or lacks NLP clarity. ``` **Task:** Rewrite the following text to adhere to strict NLP and SEO standards. **Rewriting Instructions:** 1. **Simplify Syntax:** Convert complex/passive sentences into active Subject-Verb-Object format. 2. **De-Fluff:** Ruthlessly remove *meticulous, navigating, realm, bespoke, game-changer, ever-evolving.* 3. **Humanize:** Add rhetorical asides, contractions, and direct address ("You might be wondering...") to balance the strict structure. 4. **Triple Logic:** Ensure the reader clearly understands how concepts relate (Cause -> Effect). 5. **Check Density:** Ensure the main topic is mentioned naturally (aiming for 1-2%). **Content:** [PASTE CONTENT] ``` 3.3 Featured Snippet (Position Zero) Targeter\ Goal: Win the AI Answer/Snippet box with pure clarity. ``` **Task:** Create a "Featured Snippet" definition for the question: [QUESTION]. **Formatting Rules:** 1. **Length:** 40-55 words exactly. 2. **Structure:** Start immediately with the answer. Do not use filler like "Here is the answer." 3. **Logic:** Use Semantic Triple format: [Concept] is a [Category] that [Function/Benefit]. 4. **Reading Level:** Grade 6. Zero jargon. **Question:** [PASTE QUESTION] ``` *** 4\. Technical SEO Prompts\ 4.1 Entity Schema Generator (JSON-LD)\ Goal: explicitly connect entities for Google's Knowledge Graph. ``` **Task:** Generate JSON-LD `mentions` schema for this article. Identify the top 5 entities mentioned in the text below. **For each entity, define:** 1. **The Entity Name** (e.g., "Search Engine Optimization"). 2. **The Wikipedia/Wikidata URL** (for the `sameAs` property). 3. **The Relationship:** How does this entity relate to the main topic? **Constraint:** Ensure the code is valid JSON-LD. **Content:** [PASTE CONTENT] ``` 4.2 Internal Link Opportunity Finder\ Goal: Build contextual relevance. ``` **Task:** Analyze these 5 existing article titles and summaries against my new draft. Identify 3 opportunities to link *from* the existing articles *to* the new draft. **Output:** 1. **Source Page:** (Which article?) 2. **Context:** (Where does it fit semantically?) 3. **Anchor Text:** (Create descriptive anchor text based on the destination's primary keyword—do not use "click here"). **Existing Content:** [PASTE TITLES] **New Draft:** [PASTE DRAFT] ``` *** #### 5. Editor Checklists (Updated with your Constraints and Editing SOP)
We interrupt our prompting library for an important reminder. Always check your work before publishing! AI continues to improve, however, there are still many instances where AI deviates from a quality first draft; even heavy with input (context & proper prompting). **5.1 The "2026 Editor" Checklist**\ \&#xNAN;*Run this against any draft before publishing.*
* First 25 Rule: Is the main keyword in the first 2 sentences? * Density Check: Is the keyword used 10-20 times (per 1k words)? * Banned Word Scan: Are "meticulous," "realm," and "bespoke" gone? * Syntax Test: Are sentences predominantly Subject-Verb-Object? * Triples Test: Can you identify the "Cause and Effect" in the H2s? * Humanity Test: Does it use contractions and sound like a "smart friend"? * Value Test: Does every sentence teach, feel, or entertain? *** #### 6. Strategy & Authority Building
**6.1 "Blue Ocean" Content Strategy**\ **Goal**: Find gaps competitors are ignoring.\ \*\*Also see number 12 for Authentic Engagement.
``` **Task:** I want to rank for [TOPIC], but the SERPs are crowded. Analyze the topic and propose a "Side-Door Strategy." **Thinking Process:** 1. **Long-Tail/Low Volume:** Identify variations high-DR sites ignore. 2. **Zero-Search Volume:** Identify high-intent problems that tools don't show volume for yet. 3. **Hub Concept:** Propose a unique angle that connects these smaller articles. **Topic:** [TOPIC] ``` **6.2 Competitor "Weakness" Map**\ **Goal**: Identify why the #1 ranking page is vulnerable.
``` **Task:** Analyze the top ranking page for [KEYWORD]. Don't tell me why it's good. Tell me why it's vulnerable. **Identify:** 1. **Information Gain:** What data/perspective is missing? 2. **NLP Clarity:** Is their content hard to read? (Passive voice, jargon). 3. **Entity Depth:** What related entities did they fail to mention? **Competitor URL/Content:** [PASTE URL OR TEXT] ``` *** #### 7. Advanced "Expert" Prompts
**7.1 The "Semantic Gap" Analysis**\ **Goal**: Find missing entities using NLP logic. ``` **Task:** Compare my content against the Top 3 ranking competitors (content provided below). Identify "Semantic Gaps"—terms, entities, and concepts that appear frequently in competitor content but are missing from mine. **Constraint:** Focus on "Salient Entities" (concepts central to the topic), not just random words. **My Content:** [PASTE] **Competitor Content:** [PASTE] ``` **7.2 SGE (Search Generative Experience) Snapshot**\ **Goal**: Create a text block specifically designed for AI summaries. ``` **Task:** Based on the content provided, write a 50-word "SGE Snapshot" of [TOPIC]. This must be factually dense, using simple subject-verb-object structure. No fluff. No "In this section." Just pure information designed to be parsed by an LLM. **Context:** [PASTE CONTENT] ``` *** #### **8. GEO (Generative Engine Optimization) Prompts** Based on the ProFound/GEO guidelines. These prompts prioritize "fact density" and structural formatting, which AI models prefer over fluff. **8.1 The "Fact-Density" & Citation Optimizer** **Goal:** Increase the "Information Gain" score of a section to trigger AI citations (Perplexity, SearchGPT, Google AI Overviews). ``` **Task:** Optimize the following content block for Generative Engine Optimization (GEO). **GEO Constraints:** 1. **Citation Injection:** You must include [NUMBER] specific data points, statistics, or direct quotes to support claims. (If data is not provided, use placeholders like `[INSERT STAT ABOUT X]`). 2. **Fluency Optimization:** Remove all transition words that do not add meaning (e.g., "Furthermore," "Additionally"). Keep sentences strictly Subject-Verb-Object. 3. **Entity Density:** Ensure the text explicitly names the entities (brands, tools, concepts) rather than using pronouns (e.g., use "Ahrefs" instead of "the tool"). 4. **Formatting:** Convert dense paragraphs into a Bulleted List or a Markdown Table if the data allows. AI models prefer structured data. **Current Content:** [PASTE CONTENT] **Output Goal:** A high-density, low-fluff version of the text designed to be cited as a source. ``` **8.2 The "Q\&A" Answer Engine Formatter** **Goal:** Win the "Direct Answer" slot in AI chats. ``` **Task:** Rewrite the answer to [QUESTION] using the "Inverse Pyramid" structure for GEO. **Structure Rules:** 1. **The "Direct Answer" (Sentences 1-2):** State the answer immediately. No "It depends" (unless absolutely necessary). Start with the core fact. 2. **The "Evidence" (Sentences 3-4):** Provide the "Why" or the "How" using a Semantic Triple (Entity A -> Causes -> Result). 3. **The "Nuance" (Sentence 5):** Add a specific condition or exception to show depth of knowledge. **Banned:** Do not use "In conclusion," "To summarize," or "Ultimately." **Question:** [PASTE QUESTION] ``` *** #### **9. Advanced Semantic Triples Prompts** This moves beyond simple sentence structure into logic that helps Google build its Knowledge Graph. **9.1 The Knowledge Graph "Bridge" Builder** **Goal:** Connect two distinct concepts using a logic bridge, ensuring Google understands why they are related on your page. ``` **Task:** Write a transition paragraph connecting [TOPIC A] and [TOPIC B]. **Semantic Logic Requirements:** 1. **Avoid Generic Transitions:** Do not use "Now let's talk about..." or "Moving on to..." 2. **The Triple Bridge:** The transition must use a "Cause and Effect" triple. * *Structure:* [Topic A] creates [Problem/Context] which necessitates [Topic B]. * *Example:* "Poor **Site Architecture** [Topic A] dilutes link equity, requiring a specific **Internal Linking Strategy** [Topic B] to fix indexation." 3. **Voice:** Authoritative. **Topic A:** [PASTE TOPIC A] **Topic B:** [PASTE TOPIC B] ``` **9.2 The "Entity Attribute" Audit** **Goal:** Ensure you aren't just mentioning a keyword, but defining its attributes (which Google needs for ranking). ``` **Task:** Analyze my draft for the entity: [MAIN ENTITY]. **Analysis:** Does the content explicitly mention the following "Attribute Triples"? 1. **Is-A:** (e.g., [Entity] is a [Category]). 2. **Has-Part:** (e.g., [Entity] features [Component]). 3. **Used-For:** (e.g., [Entity] is used for [Application]). **Action:** If these attributes are missing, write 3 sentences that naturally weave them into the existing text. Do not list them. Write them as prose. **Draft Content:** [PASTE CONTENT] ``` *** #### **10. Internal Linking & Anchor Text Strategy** **10.1 The Contextual Anchor Text Architect** **Goal:** Generate anchor text that describes the destination page, not the source page. ``` **Task:** I need to link to a page about [DESTINATION TOPIC] within a paragraph about [CURRENT TOPIC]. **Anchor Text Rules:** 1. **Descriptive Length:** 3-6 words. 2. **Promise of Value:** The anchor text must describe exactly what the user will get on the other side. 3. **Semantic Fit:** The sentence must make sense even if the link is removed. 4. **Variation:** Provide 3 options: * *Option A (Exact Match):* Includes the primary keyword. * *Option B (Contextual):* Describes the benefit/solution. * *Option C (The "Bridge"):* Connects the current problem to the destination solution. **Current Paragraph context:** [PASTE PARAGRAPH] ``` **10.2 The "Orphan Page" Prevention Strategy** **Goal:** When creating a new article, immediately define where it fits in the existing link graph. ``` **Task:** I am writing a new article about [NEW TOPIC]. Analyze my existing list of articles below. **Identify:** 1. **Parent Page:** Which "Pillar" page should link *down* to this new article? (Context: Hierarchy). 2. **Sibling Pages:** Which 2 related articles should link *laterally* to this new article? (Context: Related Entities). 3. **Anchor Text:** Suggest the specific sentence to add to the *existing* pages to create these links. **Existing Content List:** [PASTE LIST OF TITLES/URLS] ``` *** #### **11. Topical Maps (Hierarchy & Authority)** This is the strategic layer for "Topical Authority." **11.1 The Hierarchical Map Generator** **Goal:** Create a 3-tier structure (Hub, Sub-page, Supporting Blog) to dominate a niche. ``` **Task:** Build a Topical Map for the broad niche: [NICHE]. **Output Format (Markdown Table):** Column 1: **Tier (Level)** * *Tier 1:* The Pillar Page (Broad Guide). * *Tier 2:* The Sub-Topic (Specific Feature/Strategy). * *Tier 3:* The Supporting Content (Long-tail question/comparison). Column 2: **Title Idea** (Hook & Click style). Column 3: **User Intent** (What is the "Job to be Done"?). Column 4: **Semantic Connection** (How does this link back to the Tier 1 topic? Use Triple Logic). **Constraint:** Ensure there are at least 5 Tier 3 pages for every Tier 2 page. **Niche:** [ENTER NICHE] ``` *** #### **12. Authentic Engagement (E-E-A-T & "Experience" Optimization)** These prompts are designed to inject "Experience" into the content. Use these to rewrite Introductions, Conclusions, or "Why This Matters" sections. **12.1 The "Insider" Narrative (1st Person)** **Goal:** Establish immediate "Experience" authority by simulating a specific, visceral scenario. This is for the "40% Human Engagement" portion of your Global System. ``` **Task:** Rewrite the [SECTION NAME] using the "Insider Narrative" framework. **Voice:** First-Person ("I," "We"). **Execution Principles:** 1. **The "2 AM" Rule:** Start with a specific moment in time and space. Do not say "It is frustrating when..." Say "I was staring at the blank screen at 2 AM, realizing I had deleted the wrong database..." 2. **Sensory Details:** You must include one physical detail (e.g., "my cold coffee," "the hum of the server," "the notification ping"). 3. **Vulnerability:** Admit a mistake or a doubt. (e.g., "I honestly thought this strategy was dead..."). 4. **The "Text Message" Test:** Write this as if you are emailing a smart friend. Use contractions, fragments, and direct address. **Strict Banned Word List for this Section:** Do NOT use: *bespoke, meticulous, game-changer, seamless, robust, tapestry, unlock, realm, navigating.* **Context/Topic:** [PASTE TOPIC OR DRAFT PARAGRAPH] ``` **12.2 The "Driver's Seat" Immersion (2nd Person)** **Goal:** Force the reader to visualize the problem vividly. This is high-impact for "Pain Point" agitation. ``` **Task:** Write a "Hook" or "Problem Identification" section for [TOPIC] using Second-Person Immersion. **NLP & Style Constraints:** 1. **Subject-Verb-Object:** Keep sentences direct. 2. **No Fluff:** Remove "In the world of," "When it comes to," and "It is important to note." 3. **The "Movie Scene" Technique:** Do not describe the concept abstractly. Describe the *physical action* the user is taking when they encounter the problem. * *Bad:* "SEO is difficult." * *Good:* "You open Google Analytics, see the red arrow pointing down, and feel your stomach drop." 4. **Practical Stakes:** Connect the feeling immediately to a consequence (e.g., "That drop isn't just a number; it's your Q4 bonus disappearing"). **Topic:** [PASTE TOPIC] ``` **12.3 The "Authenticity Auditor" (Review Mode)** **Goal:** A critical agent that reviews content before you publish to ensure it passes the "Human" test. ``` **Task:** Audit the following text for "Authentic Engagement." You are a ruthless editor looking for "AI Fluff" and "Fake Empathy." **Evaluation Criteria (The Fail Conditions):** 1. **The "Generic Adjective" Scan:** Flag words like *captivating, binge-worthy, meticulous, bespoke, robust*. 2. **The Specificity Test:** Does the text mention a specific time, place, number, or name? If it is purely abstract, mark it as "FAIL." 3. **The "Smart Friend" Test:** Would a human actually say this to another human in a coffee shop? If it sounds like a brochure, mark it as "FAIL." **Output Requirement:** 1. Give the text a score (0-10) on Authenticity. 2. Identify 3 specific sentences that are "Engagement Killers." 3. Rewrite those 3 sentences to be visceral, specific, and human. **Content to Audit:** [PASTE CONTENT] ``` ***
--- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://learn.cuppa.ai/prompting-guidelines-and-library/gemini-prompt-library.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.