The Claude UGC Prompt Library

01 / Research

Customer Language Mining

Relatability is the engine of virality. And the only way to make content feel native to an audience is to use the exact language they already use to describe their own problems — not a paraphrase of it, not a polished version of it. The raw phrase. Mine it from TikTok comments, Amazon reviews, and Reddit threads before writing a single hook. Creators who skip this are guessing. Creators who do it have an unlimited supply of angles that already resonate because they came from the audience in the first place.

"The exact words people use to describe their problems are the hooks. Their language, used verbatim, makes the viewer feel seen before the product is even mentioned."

🔍 Research Prompts 5 Prompts

TikTok Comment Analyzer

Extract exact phrases, purchase-intent signals, and hook angles from raw comment data.

I'm running a UGC campaign for [PRODUCT NAME] targeting [TARGET AUDIENCE — be specific, e.g. "women 25-35 building a skincare routine on a budget"]. Below is raw comment data from TikTok videos in my niche. Read every comment and extract: 1. The 10 most frequently repeated phrases or complaints — use the exact words, do not paraphrase. 2. The top 5 recurring questions. Each question is a potential hook. 3. The 3 most emotionally charged comments. Quote them directly. 4. Every comment signalling purchase intent: "where do I get this", "is it free", "what's it called", "sending this to my sister." Quote each one. 5. The single biggest pain point in one sentence using their language, not mine. Raw comment data: [PASTE COMMENTS — minimum 30 for reliable signal, 50+ for best results]

Amazon Review Hook Extractor

Turn reviews into hooks and surface the secondary benefit that often outperforms the obvious angle.

I'm building UGC content for [PRODUCT NAME]. Below are Amazon reviews for this product or a direct competitor. Read every review and extract: 1. The 5 most specific "why I bought this" reasons — in the reviewer's exact words. 2. The 3 most vivid descriptions of the problem the product solved. Specific, visual, emotionally real. 3. Any secondary benefit that appears repeatedly — something the product does that was not the main purchase reason. This is often the best-performing campaign angle. 4. How reviewers describe themselves before finding the product. This is the hook character — the person the viewer identifies with at the opening of the video. 5. Turn the single best insight into 3 TikTok hook formats: - Story opener: "I was at [situation] when..." - Problem-reveal: "POV: you've been doing [thing] for [time] and..." - Result hook: "I [action] and now [specific change]" Amazon reviews: [PASTE REVIEWS]

Reddit Pain Point Miner

Extract trigger moments and community vocabulary that make content feel native to the algorithm.

I'm researching pain points for a UGC campaign in the [NICHE — e.g. "personal finance for people in their 20s"] space. Below is content from Reddit threads in this niche. Extract: 1. The top 5 frustrations — in the poster's exact words or close paraphrase. 2. Questions appearing more than once. These are curiosity-gap hook opportunities — the audience is actively searching for answers. 3. Specific trigger moments — the exact situations where the problem became unbearable. Quote the most vivid one directly. These are story opener hooks. 4. Any vocabulary or slang specific to this community. Using insider language signals to the algorithm which audience the content belongs to. 5. The most relatable post — the one you'd send a friend saying "this is so me." What makes it shareable in one sentence? Reddit content: [PASTE THREAD CONTENT]

ICP Builder

Synthesise all research into one reference block to paste into every scripting and hook prompt.

I've completed research on my target audience for [PRODUCT NAME]. Below is a mix of TikTok comment data, Amazon review excerpts, and Reddit observations. Build me a single Ideal Customer Profile formatted as a reference block I can paste into every future prompt. Include: 1. WHO they are: age range, situation, stage of life or problem. 2. WHAT they want most: their number one goal in one sentence. 3. WHAT they fear most: the outcome they're trying to avoid. 4. WHAT they've already tried: the alternatives that failed. This is what "I tried everything" refers to in a Mini-VSL. 5. HOW they talk: 5-8 specific phrases they use to describe their problem. Use direct quotes from the research. 6. WHAT they respond to: which emotional triggers generate the most engagement from this group — curiosity, FOMO, relatability, shock, or validation. 7. WHAT to avoid: topics, tones, or language patterns that signal non-buying younger content — stepsis jokes, rapid-cut overstimulation, child-like slang. These signals push the wrong demographic in and the right one out. Raw research: [PASTE COMBINED RESEARCH]

Secondary Benefit Finder

Surface the unexpected angles that consistently outperform the obvious primary angle.

My product is [PRODUCT NAME]. Its primary purpose is [PRIMARY BENEFIT]. Research data below from comments, reviews, and forums. Identify: 1. Secondary or unexpected use cases appearing in the research — things people use the product for beyond its main purpose. 2. Emotional benefits beyond the functional ones: confidence, less anxiety, feeling in control, social belonging. These often out-convert feature claims. 3. The "holy sh*t" moment — when people realised the product did something unexpected. What was that discovery? Quote it if possible. 4. Any angle not yet used in the niche — a framing no other creator has run with. 5. For each secondary angle found: one rough hook line (concept only, not a full script). Research data: [PASTE RESEARCH]

02 / Hooks

Hook and Angle Generation

Virality is not random. It follows a predictable chain: spark emotion → drive engagement → trigger algorithmic distribution → accumulate views. Views are the outcome of the first three steps, not the starting point. The hook is where that chain either starts or breaks. TikTok tests every video against roughly 200 users first. A video that doesn't earn a second push never gets seen. Up to 70% of that outcome is decided in the first 3–5 seconds — before most viewers have consciously decided to watch.

The Three-Part Hook — All Required

01Visual Hook

Creates involuntary attention in the first frame before the viewer consciously decides to watch. A plain wall stops nothing. Move the camera, hold the product in sunlight, drop it in water. A strange visual cannot pair with motivational text — all three components must be emotionally consistent.

02Text Hook

Functions like a newspaper headline. Not "this app helps you study" — but "I was doom scrolling at 2am the night before finals and this saved my GPA." Specific and curiosity-generating. Makes the viewer feel they would miss something by scrolling.

03Auditory Hook

Trending audio carries pre-existing emotional associations. Roughly 80-90% of top-performing audio is available on both personal and business accounts. Sound supports the hook — it does not replace the visual or text.

04Five Hook Elements

Aim for 2-3 per hook: emotional disruption (specific and concrete), curiosity gap, proof (one credibility signal), urgency (stakes + time frame), secrecy (insider framing). Two is the floor. Scoring all five in one hook is rare and usually produces the campaign's best-performing video.

The visual is often the more powerful of the two. Most creators obsess over text and neglect the first frame. Build both at the same time. All three must work together.

🎣 Hook and Angle Prompts 6 Prompts

Bulk Hook Generator

20 hooks across 5 formats with visual hook suggestions and element scores for each.

Generate 20 TikTok hooks for [PRODUCT NAME] — 4 hooks per format across 5 formats. Product: [PRODUCT NAME AND ONE-LINE DESCRIPTION] ICP: [PASTE ICP BLOCK FROM SECTION 1] Primary pain point: [EXACT LANGUAGE FROM RESEARCH — their words, not yours] Tone: [e.g. "relatable and self-deprecating" / "shocked discovery" / "direct and confident"] FORMAT 1 — Story Opener: Opens mid-situation. Specific person, specific moment. No product yet. FORMAT 2 — POV: Places the viewer directly in the scenario. "POV: you just found out..." FORMAT 3 — Regret Reveal: "I wasted [time/money] before finding this." Makes the viewer feel the pain of not knowing sooner. FORMAT 4 — Curiosity Gap: Creates the feeling of insider knowledge others do not have. FORMAT 5 — Direct Qualifier: Names exactly who the video is for. Filters everyone else out. For each hook also provide: - The visual hook: what the first frame should look like - Which of the 5 elements it activates: emotional disruption / curiosity gap / proof / urgency / secrecy

Angle Generator

8 genuinely distinct campaign angles — different framings, emotions, and sub-audiences.

Generate 8 distinct campaign angles for [PRODUCT NAME]. An angle is a specific way of framing the product — not a format. The same angle can be executed in multiple formats. Two videos can share an angle with completely different formats. Product: [PRODUCT NAME, what it does, key benefits, compliance restrictions if any] ICP: [PASTE ICP BLOCK] For each angle: 1. Name (2-4 words: "The Regret Angle" / "The Accidental Discovery Angle") 2. Hook concept in one sentence 3. Target sub-audience within the broader ICP 4. Primary emotion designed to trigger 5. Rough first-line hook example 6. How the product is positioned: solution / shortcut / insider knowledge / accidental discovery Make the 8 genuinely distinct from each other. Do not give 8 variations of the same concept with different wording.

Hook Reframer

10 reworked versions of a failing hook — one variable changed per version. For videos stuck under 500 views.

My current hook: "[PASTE EXISTING HOOK]" This video stalled under 500 views — the hook failed to earn a second algorithmic push. TikTok tested it against roughly 200 users and stopped. The problem is in the first 1-2 seconds. Do not touch the body. Product: [PRODUCT NAME] Audience: [AUDIENCE DESCRIPTION] Generate 10 reworked versions. Each must: - Keep the same underlying angle or message - Change exactly ONE variable: specificity, emotional trigger, format type (story/POV/qualifier/curiosity/claim), visual hook, or narrator character - Be under 12 words for the text portion For each: state exactly what variable changed and why it might improve scroll-stop rate. After all 10: rank 1-10 by predicted performance and explain the top 3 in one sentence each. If the original concept — not just the wording — is the actual problem, say so directly.

Viral Renaming Tool

3-7 word phrases describing the transformation, not the feature. Describe what the product does for someone's life.

My product is "[PRODUCT NAME]" and it [WHAT IT DOES IN ONE SENTENCE]. Current description: "[HOW I CURRENTLY DESCRIBE IT]" This describes what the product is, not what it does for someone's life. I need viral reframing. Examples of the format: - Skincare product: "Korean skin in a bottle" - Calorie tracking app: "a metabolism cheat code" - Budgeting app: "a broke habit detector" Give me 10 viral reframes. Each must: - Be 3-7 words maximum - Describe the outcome or transformation, not the feature or category - Use vocabulary the target audience actually uses - Be memorable enough to say out loud in a video without sounding scripted - Pass this test: would hearing it make someone immediately want to know more? Target audience: [AUDIENCE DESCRIPTION or PASTE ICP]

Stitch Hook Generator

Counter-takes and authority reactions using TikTok's Stitch feature. Two proven formats with first-line scripts.

I want to use TikTok's Stitch feature — opening with a clip from another creator, then responding in a way that naturally leads to my product. Product: [PRODUCT NAME AND WHAT IT DOES] Audience: [AUDIENCE DESCRIPTION] Two Stitch formats: - Counter take: stitch a video making a claim, disagree or correct it, close with the product as the better solution - Authority reaction: position as an expert reacting to another creator's content, product surfaces as the professional recommendation Generate: 1. 5 counter-take hook concepts: what kind of claim would I stitch, and what is my counter-position? The product should be the natural conclusion of the disagreement. 2. 5 authority reaction hook concepts: what kind of content would I react to, and what expert framing would I use? 3. For each of the 10: write the first spoken line I say after the stitched clip ends (under 15 words). 4. What visual hook should I use in my own footage to reinforce the text and audio?

Niche Borrow Tool

Adapt a high-performing format from another niche. The core emotional mechanic stays. Context changes.

I found a content format performing well in a different niche. A format that works in one niche can usually be translated — the core emotional mechanic stays the same; the product and context change. Original format (different niche): [DESCRIBE IN DETAIL: hook structure, what happens in the body, how it ends, what emotion it triggers, why it works, view counts if known] My product: [PRODUCT NAME AND DESCRIPTION] My audience: [AUDIENCE DESCRIPTION or PASTE ICP] Adapt this format. Provide: 1. Translated hook concept — same emotional mechanic, different context 2. Rough script outline following the same structural beats 3. What my visual hook should look like in this context 4. Any elements from the original that will not translate cleanly, and what to replace them with 5. One risk: what could go wrong if the product is a poor fit for this format's emotional logic?

03 / Scripts

Script Writing — All 4 Video Formats

Views do not make money. Conversions do. A video with 100 million views that drives no purchases is worthless. The difference between a video that goes viral and converts versus one that goes viral and produces nothing comes down to one question: if the product were removed from the video, would the video still make sense? If yes, it will not convert. The product must be load-bearing — the reason the story exists, not something bolted on at the end. All four video formats and four scripting frameworks are here. The format is how the video looks. The framework is the persuasion structure underneath.

Product Not Load-Bearing

Remove the product and the video still makes complete sense. The product was bolted on at the end. Goes viral for the wrong reason — views without revenue. The test: remove the product. If the story survives intact, it will not convert.

Product Load-Bearing

Remove the product and the story collapses — no punchline, no resolution, no release of tension. The product is the reason the video exists. Run this test on every script before filming. This is the structure that converts.

✍️ Script Prompts 9 Prompts

Talking Head — Problem-Solution

Direct-to-camera. Most versatile framework. Works for any product with a clear pain point.

Write a 30-45 second talking head TikTok script using the Problem-Solution framework. Talking head = direct-to-camera delivery. Before writing: define "My video helps [person] do [thing] so they can get [benefit] without [pain point]." Every line must connect back to this anchor. The setting signals authority before a word is spoken — film where the target audience would be. Product: [PRODUCT NAME, what it does, key benefits, compliance restrictions] ICP: [PASTE ICP BLOCK] Hook: [PASTE CHOSEN HOOK] Structure: 1. HOOK / PROBLEM (3-5 sec): Open with the hook. Target audience must recognise themselves immediately. 2. DEEPEN THE PROBLEM (5-10 sec): Failed alternatives or escalating stakes. No product yet — desire must be built before the reveal. 3. PRODUCT REVEAL (5-8 sec): Introduce naturally — like relief, not an ad. Do not name the product before the 15-second mark. Naming it early gives viewers a reason to leave TikTok and search directly, which kills watch time. 4. DEMONSTRATION (8-12 sec): One feature at a time. Specific, visible progress. No feature dumps. 5. CTA (2-3 sec): "Link in bio" or the product name. Nothing more. Format rules: on-screen text 5 words max per clip, body clips 1-3 seconds. Write in spoken casual voice. Include [VISUAL CUE] notes for the creator.

Reaction and Demo Format

Visual reaction hook, then product walkthrough. Pattern interrupt required in the first frame.

Write a 25-40 second reaction-and-demo TikTok script for [PRODUCT NAME]. Reaction and demo opens with a visual reaction clip then transitions into a product walkthrough. The text hook functions like a newspaper headline. The visual hook must be a pattern interrupt — move the camera, cover the mouth, zoom in. Static frames do not stop scrolls. Do not name the product until the final seconds. Product: [PRODUCT NAME, key features to demonstrate, compliance restrictions] ICP: [PASTE ICP BLOCK] Structure: 1. VISUAL REACTION HOOK (0-3 sec): A pattern interrupt. What does the creator do that stops the scroll before any words? 2. TEXT HOOK ON SCREEN: Signal value and create curiosity before the viewer hears anything. 3. PROBLEM STATEMENT (3-8 sec): Surface the problem. No product yet. 4. PRODUCT WALKTHROUGH (10-20 sec): Problem → action in product → result → optional bonus feature. Body clips 1-3 seconds. On-screen text 5 words max. 5. PAYOFF LINE (final 3 sec): "How did I not know about this sooner?" or "App is called [name]." Product name lands last or not at all. Include [VISUAL CUE] notes throughout.

Skit Format

Partially improvised — concept over script. Product must be load-bearing or the format fails entirely.

Write a skit concept and loose scene outline for [PRODUCT NAME]. Skits are entirely reliant on relatability. The product must be load-bearing: remove it and the skit collapses. If the skit can exist without the product, the integration is too weak. Avoid heavy scripting — partially improvised delivery with a clear concept outperforms word-for-word scripts. Find skit formats by browsing the For You page with intention, then ask: can the product fit naturally? Product: [PRODUCT NAME AND WHAT IT DOES] ICP: [PASTE ICP BLOCK] Give me: 1. 3 distinct skit concepts — different scenarios where the product naturally becomes load-bearing 2. For the strongest concept: a loose scene outline with beats (not word-for-word) — what happens, in what order, where the product appears, how it ends 3. Visual hook for the first frame of each concept 4. Improvisation guidance — what the creator can vary freely, and what must stay consistent for the product to land 5. The one line the creator must not skip — where the product becomes the payoff Load-bearing test: if the product were removed from the scene, would the skit still make sense? If yes, that concept does not pass. Give me ones that fail this test.

Carousel / Slideshow — All 3 Sub-Formats

Tips list, punchline, and trend jack. Three non-negotiables: strong first slide, visual congruency, matched audio.

Write carousel content for [PRODUCT NAME] across all 3 carousel sub-formats. Three non-negotiables for every carousel: (1) strong first slide — same job as a video hook, (2) full visual congruency across every slide — same setting, tone, style, aspect ratio, (3) audio that matches the intended emotion exactly. Product: [PRODUCT NAME AND WHAT IT DOES] Audience: [AUDIENCE DESCRIPTION or PASTE ICP] FORMAT 1 — TIPS LIST: Hook slide signals value. Middle slides deliver genuine tips the audience would share even without the product. One slide introduces the product as a recommendation. Non-CTA slides must deliver real standalone value — the list should be worth saving even without the product mention. Write: hook slide text + 4-6 tip slides + product recommendation slide FORMAT 2 — PUNCHLINE: 2-3 slides total. Setup then punchline. The punchline IS the product. Better for reach than conversion. After posting, pin a comment pointing to the product — this is part of the format, not optional. Write: setup slide(s) + punchline slide + what the pinned comment should say FORMAT 3 — TREND JACK: Adapt a currently trending carousel format to the product. Trend jacks have short shelf lives — use while the format is active. Write: describe a current trending carousel structure, then show how to adapt it with the product as the subject

Hook → Fear → Proof → Reveal

Conversion-optimised. Desire built before the product appears. For identity and risk-based problems.

Write a 30-50 second TikTok script using the Hook to Fear to Proof to Reveal framework. This is conversion-optimised. The product does not appear until the viewer emotionally wants the solution. Do not reveal it early. Product: [PRODUCT NAME, key benefits, compliance restrictions] ICP: [PASTE ICP BLOCK] Hook: [PASTE CHOSEN HOOK] Structure: 1. HOOK (3-5 sec): Stop the scroll. Surface an existing fear — aspiration, identity, risk, or emotional disruption. Do not manufacture new fears. Bring existing ones to the surface. 2. FEAR (5-8 sec): Name the problem in 2-4 lines using exact language from the ICP research. The viewer should think "this is describing me." 3. PROOF (3-5 sec): One line. One specific result. Not a list — one thing. "I went from X to Y in Z time." This comes before the reveal. "I've been using this for a while" is not a proof line. 4. REVEAL (10-15 sec): Introduce the product only after desire is built. Sell the transformation, not the features. Write in first-person spoken voice. Include [VISUAL CUE] notes.

Mini-VSL (Intriguing Storyline)

40-60 seconds. Story builds before the product appears. Includes the time-jump mechanic and anti-establishment close.

Write a 40-60 second TikTok script using the Intriguing Storyline (Mini-VSL) framework. For products that need emotional investment before the viewer is ready to receive the pitch. A story builds before the product appears. Runs 30-60 seconds. Product: [PRODUCT NAME, what it does, key benefits, compliance restrictions] ICP: [PASTE ICP BLOCK] Hook: [PASTE CHOSEN HOOK] Structure: 1. BROAD RELATABLE HOOK (3-5 sec): An emotion or situation the entire target audience feels. Not product-specific yet. 2. DEEPEN THE ISSUE (5-10 sec): Specific details. The viewer thinks "this is exactly me." 3. FAILED SOLUTIONS (5-8 sec): "I tried everything and nothing worked." Name 2-3 specific alternatives that failed. This validates the viewer's experience of having already tried to solve this. 4. TIME JUMP: Use the exact phrase "fast forward [X] weeks later" — this is a deliberate retention mechanic. It creates anticipation and signals a dramatic reveal is coming. 5. PRODUCT REVEAL (8-12 sec): Specific improvements only. No generic claims. 6. FEAR OR SCARCITY (3-5 sec): "I almost didn't try it" (fear of inaction) or "it keeps selling out" (manufactured scarcity). 7. ANTI-ESTABLISHMENT CLOSE (optional, 3-5 sec): "Big [industry] doesn't want you to know about this." Triggers reactance bias — the product becomes insider knowledge, not an ad, elevating perceived value and creating urgency. 8. CTA (2-3 sec): Product name or "link in bio." Include [VISUAL CUE] notes. Spoken casual voice throughout.

Authority Template (Dentist / Psychiatrist)

5-block structure. Bypasses scepticism via layered credibility. Discovery must feel accidental and non-commercial.

Write a 45-60 second TikTok script using the Authority Template (Dentist / Psychiatrist structure). This bypasses viewer scepticism by routing credibility through an established authority figure before the product appears. The discovery must feel accidental and non-commercial. Product: [PRODUCT NAME, scientific or clinical ingredients if relevant, what it does, compliance restrictions] Authority figure: [WHO — e.g. "my father, a dentist of 25 years" / "my therapist" / "a nutritionist I know" — make it personal and specific] Structure: 1. AUTHORITY SETUP (0-20 sec): Introduce the figure. Establish credentials. Introduce a serious problem they could not solve. Include their initial scepticism. 2. TENSION AND TIME GAP (5-8 sec): Failed traditional approaches. Then: "Fast forward [X] months." Deliberate retention mechanic — creates anticipation. 3. DRAMATIC REVEAL (8-10 sec): The authority figure expresses genuine shock at the result. Specific improvements only — no vague claims. Their amazement is the social proof. 4. ACCIDENTAL DISCOVERY (8-10 sec): The creator researched specific clinical ingredients, not a product — found this by accident. Makes the find feel non-commercial, bypassing the viewer's ad defence. 5. SCARCITY CTA (5-8 sec): Personal testimony then manufactured scarcity. "The only problem is it's always selling out." Specific instruction: "If you see [visual signal], it's in stock. Buy it now." Write in casual first-person. Include [VISUAL CUE] notes.

Script Load-Bearing Audit

Catch integration failure, dead air, and early payoff before filming. Run on every script.

Audit this script for [PRODUCT NAME] before it goes to my creator. 1. LOAD-BEARING TEST: If I removed every mention of the product, would the video still make sense? Yes or no — explain why. 2. DESIRE TIMING: Is the product revealed before or after the viewer has had a chance to emotionally want the solution? If revealed too early, identify the exact line where the reveal should be delayed to. 3. DEAD AIR SCAN: Identify every line or section that does not advance the story or build toward the reveal. Mark each as "cut" or "tighten." 4. PAYOFF PLACEMENT: Where does the payoff land? Is it in the final 20% of the video? If not, show how to restructure. 5. OPEN LOOP CHECK: What question or tension does the hook create that keeps the viewer watching? If you cannot identify a clear open loop, flag it — the video likely has a retention problem regardless of hook quality. 6. CTA AUDIT: Is the CTA earned? Has the viewer received enough value by the time the CTA appears? 7. ONE-LINE FIX: The single most important change before filming. Script: [PASTE FULL SCRIPT]

Controversy and Engagement Layer

Add a deliberate engagement trigger to any finished script. Product-related controversy only. Target: curiosity and surprise, not anger.

I have a finished script for [PRODUCT NAME] and want to add a deliberate engagement trigger. Right kind of controversy: product-related — a bold falsifiable claim about the niche, or something confusing about how the product works. Wrong kind: controversy targeting any group (damages the account long term regardless of view count), or anger-farming (brings in non-buyers and distorts the audience pool). Target emotions: curiosity, relatability, surprise, mild shock. Not anger. Existing script: [PASTE SCRIPT] Suggest 3 triggers — one from each type: 1. CONFUSION TYPE: Something that does not quite add up, causing viewers to rewatch or jump to comments. Specific placement and specific element. 2. BOLD CLAIM TYPE: A specific, falsifiable claim about the product or niche that invites debate. Must be product-related and defensible. Exact line and placement. 3. VISUAL ABSURDITY TYPE: Something unexpected in the foreground or background — unrelated to the product — that generates comment activity. For each: placement in the script, what it is, and whether the comment activity it generates is likely to attract the buying demographic.

04 / Creators

Creator Outreach and Delivery Coaching

A mass UGC system is an owned asset, not a rental. The goal is not to find expensive influencers with bloated prices — it's to find undiscovered talent who possess raw potential but haven't been commoditised yet. Two creators delivering the exact same script can produce significantly different results based on delivery alone. When one account consistently outperforms another on the same angle and script, the problem is almost never the script.

The 4 Delivery Variables That Separate Accounts

01Speaking Cadence

Fast, energetic delivery with a controlled rhythm. Short pauses between sentences. No flat or drawn-out phrasing. A viewer's brain has idle capacity when the speaker is slow — idle capacity becomes the impulse to scroll. Fast delivery keeps the brain working to keep up, which keeps the viewer on screen.

02Eye Contact

Look directly into the camera throughout. Looking away, looking down at notes, or drifting to the side signals low confidence and breaks the viewer's sense of being spoken to directly. One of the simplest delivery variables to fix and one of the most consistently overlooked across campaigns.

03Secondary Foreground Action

A creator cutting fruit, doing makeup, or cooking while talking holds attention more effectively than a static creator. The physical action gives the viewer's eye something to track while the ear processes the message. Both sensory channels are occupied. This is not the same as background clutter — a foreground action is intentional. Background clutter is an attention drain.

04Setting Congruence

The setting should match the target audience's world. A study app creator should be at a desk. A fitness app creator should be in a gym. The setting signals credibility before a word is spoken. A mismatched setting — a financial app creator filming in a bedroom with gaming posters — erodes trust silently and invisibly.

When two accounts run the same angle but one consistently outperforms the other, compare delivery cadence and secondary foreground action before touching the script. Delivery is often the only variable that differs.

🎞️ Creator Prompts 4 Prompts

Creator Outreach DM Generator

Sourcing DMs, cold email subject lines, and screening questions. Target: 2k-10k followers, 1-5 videos with 100k+ views.

Write outreach messages to source UGC creators for [PRODUCT NAME / BRAND]. Target creator profile: 2,000-10,000 followers, 1-5 videos with 100k+ views, shows face on camera, US-based, not currently running a brand program (no "contact my manager" in bio, no obvious brand deals). Avoid creators already working for AI startups, large UGC programs, or anyone who appears to be a program lead — leads are paid 4-5x the standard rate. Campaign details: - Base pay: [MONTHLY RATE — e.g. "$500/month base plus view bonuses"] - Posting frequency: [e.g. "5-7 posts/week"] - View bonuses: [e.g. "$100 per 100k views" or "$250 per 250k views"] - Account setup: new account created for this campaign, not their main page - Time commitment: low Generate: 1. TikTok / Instagram DM (under 50 words): Lead with the paid collaboration. Template that works: "[paid collab] Love your [specific content reference]. Would love to collab. We pay well, low time commitment, and this wouldn't be on your main page. Lmk if you're interested!" Casual — not corporate. 2. Cold email subject lines (5 options): The dollar amount is what gets the email opened. Lead with it. Format: "$[AMOUNT]/month PAID collab." 3. Application form screening questions (5 questions): Filter for camera charisma, aesthetic sense, and natural storytelling — the three traits that matter — without relying on follower counts.

Creator Delivery Audit

Score a creator against the 4 delivery variables. Generate specific coaching notes ready to send via WhatsApp.

Audit a creator's delivery and generate coaching notes before their next filming session. Creator performance: - Average views (last 5 posts): [VIEWS] - Engagement rate: [RATE] - Account age and warm-up status: [DETAILS] Delivery observations (describe what you see in their videos): - Speaking cadence: [FAST / SLOW / FLAT / ENERGETIC — describe specifically] - Eye contact: [DIRECT THROUGHOUT / DRIFTS / LOOKS DOWN — describe] - Secondary foreground action: [YES — describe what / NO — static] - Background: [CLEAN / CLUTTERED — describe] - Setting congruence: [Does the setting match the target audience's world?] Comparison (if available): - Creator B average views on the same angle: [VIEWS or N/A] - What Creator B does differently in delivery: [DESCRIBE or N/A] Generate: 1. Delivery score for each of the 4 variables (1-5 scale with one sentence explanation) 2. The single most impactful change for the next 5 videos 3. Word-for-word coaching notes to send via WhatsApp — casual, specific, direct 4. One specific filming instruction addressing the biggest gap: e.g. "Film while cooking something in the foreground" / "Record at your desk with your laptop open behind you"

Comment Seeding Generator

Organic-looking planted comments and creator reply scripts. Act immediately at 10k+ views.

My video for [PRODUCT NAME] is gaining views. I need organic-looking comments that signal demand and give a natural reason to reply with the product name. The planted comment comes from a separate account that looks like a real viewer. The creator replies immediately: simple, under 10 words, no enthusiasm or feature explanation. "it's called [name] btw" is the right energy. Respond to the first 30 comments as fast as possible — the algorithm pays extra attention to that first wave of activity. Product: [PRODUCT NAME] Video topic: [WHAT THE VIDEO IS ABOUT IN 1-2 SENTENCES] Current views: [VIEWS — act immediately at 10k+] Generate: 1. 8 organic-looking planted comments: vary phrasing and tone so they don't look coordinated. Include: "what app is this?" variations, "is this free?" variations, sceptical "does this actually work" variations, and "sending this to [person]" social-sharing variations. 2. 5 creator reply scripts: under 10 words each. Casual. Names the product without sounding like an ad. No enthusiasm, no feature explanations. 3. Drop comment (if no planted comment arrives): what to post from the creator account if the video crosses 10k+ with no organic product-intent questions yet. 4. 3 video reply openers: if making a video reply to one of these comments, what is the first spoken line?

Viral Format Replication Brief

The message that goes to every creator the moment a format breaks out. Viral formats have a short window — move immediately.

A video just broke out. I need a replication brief for every creator immediately. When a video goes viral, move fast — have every creator copy the format word for word, second by second, before it goes stale or a competitor runs it into the ground. Winning video: [DESCRIBE IN FULL: hook text, visual hook, format type, script summary second-by-second, timing, what the creator was doing physically, background, audio, what specifically made it different from other videos] Performance: [VIEWS, ENGAGEMENT RATE, NOTABLE COMMENTS] Write a creator brief that: 1. Opens with the performance numbers — creators need to understand why this matters 2. Gives a second-by-second breakdown of what to recreate — visual, text, audio, delivery 3. Lists exactly 3 things they must NOT change (the variables that drove performance) 4. Lists exactly 3 things they CAN vary (so multiple accounts don't post identical content) 5. Specific delivery coaching: cadence, eye contact, secondary foreground action, background 6. Hard deadline and posting instructions Tone: urgent, direct, excited. Write it like a WhatsApp message, not a strategy document.

05 / Diagnostics

Video and Account Diagnostics

Every underperforming video fails on exactly one of three axes — distribution, engagement, or conversion. They are independent problems with independent fixes. Combining them into a single diagnosis produces a single wrong answer. The most common mistake is optimising the script when the problem is the account, or fixing the hook when the problem is that the product is not load-bearing. Always check account health before touching content. You cannot script your way out of a flagged account.

<500

Views = Hook Failure

500–3K

Views = Retention Issue

<300

Views = Account Problem

48 hrs

Min Age to Diagnose

🩹 Diagnostic Prompts 7 Prompts

Full Three-Axis Video Diagnosis

Distribution, engagement, conversion — in order. One fix per axis. One output at the end.

Diagnose this video across three independent axes. Run them in order. Do not combine them. First check: is this video at least 48 hours old? If not, flag it and note current trajectory — do not diagnose yet. Views can spike late and early data is unstable. Video data: - View count: [VIEWS] - Engagement rate (likes + comments + shares / views): [RATE] - Likes / Comments / Shares / Saves: [COUNTS] - Age of video: [HOURS OR DAYS] - Account age (date of first post): [DATE] - Average views on last 6 posts: [AVERAGE] - Age of the angle (date first used): [DATE] - Shadowban or flag recently: [YES/NO] - Hook and video summary: [DESCRIBE] - Product-intent comments: [YES/NO — quote examples if yes] AXIS 1 — DISTRIBUTION: Where did it stall? Is most of the last 6 posts stuck under 300 views? If yes: account problem, not content problem — stop here and use the Account Health Check. If no: where in the distribution curve did this video stop? AXIS 2 — ENGAGEMENT: Is the content emotionally flat? What does the engagement rate and comment-to-like ratio indicate? AXIS 3 — CONVERSION: Only run if 3,000+ views AND 2%+ engagement. Is the product disconnected from the content? Final output: one specific thing to fix. Not three. One. If the angle is over 4 weeks old with declining views, flag saturation first — this is an angle age problem, not a content quality problem.

Account Health Check

Run this before any content diagnosis. Outputs one of four states with a specific next action.

Run an account-level diagnosis before I touch any content. Account problems and content problems require completely different fixes. Account data: - Date of first post: [DATE] - Total posts: [NUMBER] - Average views — first 10 posts: [AVERAGE] - Average views — most recent 10 posts: [AVERAGE] - Posts stuck under 300 views in last 6: [HOW MANY] - Warm-up completed (5+ days, 1hr/day engaging niche content before first branded post): [YES/NO] - Posts per week: [NUMBER] - Giveaways run on this account: [YES/NO] - Shadowbans or guideline strikes: [YES/NO — details if yes] - Account type (personal or business): [TYPE] - US SIM + dedicated US VPN (for non-US posting): [YES/NO/PARTIAL — has VPN ever lapsed?] - Geographic breakdown (where are views coming from): [COUNTRIES AND % IF KNOWN] Return one of four states with a specific next action: 1. HEALTHY — distribution building, focus on content quality 2. WARMING UP — too early to diagnose, confirm warm-up compliance 3. FLAGGED / DAMAGED — likely suppressed; stop branded posting for 1-2 days, scroll For You page like a normal user, do not go completely silent, resume only after views recover 4. AUDIENCE-POISONED — wrong follower pool from giveaways or off-target viral content; usually faster to start a new account than to rehabilitate

Geographic Targeting Check

Catch distribution drift. High views from the wrong countries mean zero conversions regardless of content quality.

I have a TikTok account getting views but unexpectedly low conversions. Diagnose whether the algorithm is pushing content to the wrong geographic pool. A video can generate millions of views from regions the brand does not serve. This is a distribution problem, not a content problem. Geographic pools are determined by IP address, network carrier, and device timezone. A single VPN lapse — even once — can reset the targeting profile. Account data: - Primary target market: [COUNTRY / REGION — e.g. "United States"] - US SIM card installed in device: [YES/NO] - Dedicated US VPN active before opening TikTok: [YES/NO — has it ever lapsed, even once?] - Account created on this device with VPN active: [YES/NO] - Geographic breakdown from TikTok analytics (last 30 days): [LIST COUNTRIES AND % OF VIEWS] - Conversion data: [DOWNLOADS / INSTALLS / PURCHASES — are they from the target country?] Diagnose: 1. Is geographic distribution aligned with the target market? 2. Could a VPN lapse or network carrier mismatch explain the drift? 3. If geographic drift is confirmed: can this account be salvaged, or should a new account be created from scratch with the full warm-up protocol and correct device setup? 4. What posting time in the target timezone should this account use going forward?

Winning Video Reverse Engineer

Break down exactly what drove a breakout video. Replicate immediately across every account.

A video significantly outperformed everything else on this account. Identify exactly what drove it so I can replicate immediately. Video data: - View count: [VIEWS] - Engagement rate: [RATE] - Notable comments (especially product-intent): [QUOTE EXAMPLES] - Shares and saves: [COUNTS] - Account age at time of posting: [ACCOUNT AGE] - Angle age: [HOW LONG THIS CONCEPT HAD BEEN USED] Hook and video summary: [DESCRIBE: hook text, format, what happened, visual hook, delivery, setting, audio] What was different from other videos on this account: [WHATEVER CHANGED: wording, visual, format, delivery, secondary foreground action, setting] Analyse: 1. Which specific element most likely earned the second algorithmic push? 2. The replicable structure to apply to the next 5 videos on this account 3. The one variable NOT to change in the next iteration 4. The one small test to run in the next video while keeping the core structure intact 5. Which other creators in the program should receive this format immediately, and in what order?

Creator Comparison Diagnosis

Isolate the variable explaining a performance gap between two creators on the same script.

Two creators are running the same angle and script for [PRODUCT NAME] with significantly different results. Isolate the primary variable. Creator A: - Average views (last 5 posts on this angle): [VIEWS] - Account age and warm-up status: [DETAILS] - Angle age on this account: [DATE STARTED] - Delivery: [CADENCE, EYE CONTACT, SECONDARY FOREGROUND ACTION, BACKGROUND — describe specifically] Creator B: - Average views (last 5 posts on this angle): [VIEWS] - Account age and warm-up status: [DETAILS] - Angle age on this account: [DATE STARTED] - Delivery: [CADENCE, EYE CONTACT, SECONDARY FOREGROUND ACTION, BACKGROUND — describe specifically] Check variables in this priority order: 1. Structural: account age difference, warm-up compliance, audience pool, geographic targeting 2. Delivery: speaking cadence, eye contact, secondary foreground action, background cleanliness 3. Hook execution: text wording specificity, visual hook strength, emotional consistency 4. Product integration: is the product more load-bearing on one account than the other? Output: primary variable, confidence level (high / medium / low), and a single-variable test for the underperforming creator to run over the next 3-5 videos.

Video Comparison — 4 Types

Identify the comparison type first. The type determines which variables are meaningful to check.

Compare two videos and isolate the variable explaining the performance gap. The comparison type determines which variables to check. Step 1: Identify the comparison type before doing anything else: - TYPE 1: Same account, same angle — execution quality is the primary variable (most controlled comparison) - TYPE 2: Same account, different angles — angle concept is the primary variable - TYPE 3: Different accounts, same angle — delivery and account health are the primary variables - TYPE 4: Different accounts, different angles — directional only; cannot cleanly isolate variables Video A: - View count / Engagement rate: [STATS] - Account age / Angle age on this account: [DATES] - Hook text (exact wording): [TEXT] - Visual hook (first frame): [DESCRIBE] - Delivery: [CADENCE, EYE CONTACT, FOREGROUND ACTION, BACKGROUND] Video B: - View count / Engagement rate: [STATS] - Account age / Angle age on this account: [DATES] - Hook text (exact wording): [TEXT] - Visual hook (first frame): [DESCRIBE] - Delivery: [CADENCE, EYE CONTACT, FOREGROUND ACTION, BACKGROUND] Check in this priority order based on the comparison type: 1. Account health (for Type 3 and 4: check both accounts with F2 before attributing gap to content) 2. Delivery quality (cadence, eye contact, secondary foreground action, background) 3. Hook quality (text wording, visual hook, emotional consistency) 4. Script structure (open loop, payoff timing, product integration) Output: comparison type, primary variable, confidence level, specific next test.

Angle Saturation Check

Is it the angle saturating or just the hook wording? These require different fixes — do not confuse them.

Assess whether this angle is saturating or just needs a hook refresh. These are different problems with different fixes. Angle description: [DESCRIBE THE CENTRAL HOOK CONCEPT IN 2-3 SENTENCES] Date first used on this account: [DATE] Total videos posted with this angle: [NUMBER] Performance by week (oldest to most recent): [e.g. "Week 1: 12k, 8k, 15k / Week 2: 9k, 7k, 6k / Week 3: 4k, 3k, 3k"] Comments on recent videos saying they've seen this: [YES/NO — quote examples if yes] Diagnose: 1. Is the view trend ascending, flat, or declining? How many consecutive videos show decline? 2. Is the core concept still valid, or is it specifically the hook wording and visual execution that has gone stale? These are different problems. 3. If concept still valid: write 3 hook refresh options — different opening lines, same underlying angle. Do not abandon the structure. 4. If fully saturated (4+ weeks, consistent decline, comments saying "seen this"): generate 3 new angle hypotheses using secondary product benefits, comment-section observations from the best-performing videos, or adjacent ICP problems. 5. Decision: keep running with hook refresh / retire / pivot — with specific reasoning from the data.

06 / Scaling

Angle Variation, Scaling, and Paid Social

Getting views in month one is different from still getting views in month six. Longevity comes from finding a format that works and building a system to keep it fresh — not abandoning it the moment it performs. When a format goes viral, move immediately. Have every creator copy it word for word before it goes stale. The second lever is paid: your top-performing organic videos are battle-tested creatives validated by the most ruthless algorithm in the world. They become the primary fuel for Spark Ads and Meta campaigns — cutting CPI and lifting conversion rates beyond what standard ad creatives can achieve.

"When a video goes viral, move immediately. Have every creator copy the format word for word, second by second, before it goes stale or a competitor brand runs it into the ground."

⚡ Scaling + Paid Social Prompts 6 Prompts

50 Variation Generator

One winning format into 50 testable variations. One variable changed per variation — never two.

I have a winning format for [PRODUCT NAME] and need 50 variations to run across multiple creators and accounts. Winning format: [DESCRIBE IN FULL: hook text, visual hook, script structure, format type, delivery notes, what made it perform] Performance: [VIEWS AND ENGAGEMENT RATE] Rule: change exactly ONE variable per variation. Never two at once — you need to know what moved the needle. HOOK WORDING VARIATIONS (10): Same concept, different opening line. Vary: the character narrating, specificity of the scenario, which hook element leads. VISUAL HOOK VARIATIONS (10): Same script, different first frame. Vary: camera angle, action, setting element, camera movement. AUDIENCE TARGETING VARIATIONS (10): Same hook, adjusted for a slightly different sub-segment of the ICP. Who else has this problem in a different context? FORMAT VARIATIONS (10): Same angle, different format — talking head, reaction-demo, skit, carousel. The core message stays identical. CTA AND ENDING VARIATIONS (10): Same opening and body, different final 5 seconds. Vary: CTA placement, wording, scarcity element, whether to name the product or not. Label each: [Axis] — [Number] — [What changed and why it might perform differently]

4-Format Rotation System

Build a 90-day format library. One format saturation cannot kill the account if 3 others are carrying it.

Design a 4-format rotation system for [PRODUCT NAME] that sustains an account for 90 days. An account relying on a single format is one bad month away from dying. The goal is 2-4 proven formats rotating so that when one loses momentum, the others carry the account while a new primary is found. Campaign context: [PRODUCT NAME, target audience, primary pain points, content restrictions, current best-performing angle if known] Posts per week per account: [NUMBER] Number of creators: [NUMBER] For each of the 4 formats: 1. Format name and type (talking head / skit / reaction-demo / carousel) 2. The hook concept anchoring it 3. Emotion designed to trigger 4. Target sub-audience within the broader ICP 5. Posts per week for this format 6. Built-in novelty variable — what changes with each iteration so the format never runs dry 7. Early warning signs this format is saturating Include a suggested weekly posting schedule distributing all 4 formats across the week.

Spark Ad Readiness Evaluator

Which organic videos are ready to boost, which need a splice, which to skip. Ranked shortlist with rationale.

Evaluate which organic TikTok videos for [PRODUCT NAME] should be run as Spark Ads. Spark Ads boost an organic video directly from the creator's profile. The organic version must run as-is — it cannot be edited before boosting. Likes, comments, and shares stay on the original post, building social proof and typically lowering acquisition costs. Ad thresholds: 5k-50k views with 7-10%+ engagement = strong signal. 50k+ views with 5%+ engagement = excellent. Under 3% engagement at any volume = weak signal, do not boost. Comment signals to look for: "What app is this?", "Is it free?", "Where do I get this?" When product-intent comments, high shares, and high saves are all present simultaneously — test as an ad immediately. Video library: [FOR EACH VIDEO: view count, engagement rate, like-to-view ratio, shares, saves, age, notable product-intent comments] Landing page / app store: [WHERE THE AD SENDS PEOPLE — describe briefly] For each video: 1. Does it meet the engagement threshold for ad testing? 2. PRODUCT CLARITY TEST: By the end of the video, does a first-time viewer understand what the product is and what it does? Yes or no. 3. CONGRUENCE TEST: Does the video's message match the landing page experience? A mismatch kills conversions at the final step. 4. AD READINESS: Ready to boost as-is / needs a splice (pull the hook, add demo after) / not suitable. 5. If splice needed: describe exactly what to add and where. Ranked shortlist of top 3 to test first with rationale.

Meta Partnership Ad Selector

For paid social. Identifies which organic UGC to whitelist on Meta, how to set up dark posts, and which demographic each video is suited for.

I'm running paid social for [PRODUCT NAME] and want to identify which organic UGC videos to run as Meta Partnership Ads (formerly whitelisted / branded content ads). Meta Partnership Ads differ from TikTok Spark Ads in two critical ways: (1) they use account-level permissions, not per-video codes, so once a creator approves access the brand can boost any of their posts or upload new variations, (2) they can appear from the creator's handle without showing on their organic profile (dark posts), which means you can test new video variations without polluting the creator's feed. Meta website campaigns reach 35-55+ audiences with higher purchasing power. TikTok app promotion campaigns skew 25-34. If the product sells to an older demographic with disposable income, Meta Partnership Ads are the primary paid channel, not Spark Ads. Creator accounts for consideration: [FOR EACH CREATOR: handle, top 3 videos with view count and engagement rate, demographic of their audience if known] Product landing page / app store URL: [URL AND BRIEF DESCRIPTION] Target buyer demographic: [AGE RANGE AND DESCRIPTION] Campaign objective: [WEBSITE TRAFFIC / APP INSTALLS / PURCHASES] Evaluate and recommend: 1. CHANNEL DECISION: For this product and demographic, should the primary paid spend go through TikTok Spark Ads or Meta Partnership Ads — or both? Explain based on the demographic data. 2. VIDEO SELECTION: Which 3 videos from the creator library are strongest candidates for Meta whitelisting? For each: engagement rate, whether a first-time viewer understands the product, and whether the message would survive landing on a 35-55 year old's Instagram feed. 3. DARK POST OPPORTUNITIES: Are there any angles or hooks we should create specifically as dark posts — content that would not work on the creator's organic profile but would perform as paid social to an older demographic? 4. SETUP CHECKLIST: What steps does the creator need to take inside their Instagram app to grant Partnership Ad access? What does the brand do inside Meta Ads Manager to activate it? 5. MESSAGE MATCH TEST: Does the video's specific claim match what the landing page delivers? A mismatch at this stage is the most common conversion killer in paid social.

Campaign Status Report

Raw data in, prioritised action plan out. Three actions maximum — not five, not ten.

Generate a campaign status report and prioritised action plan. Campaign data: - Start date: [DATE] - Active accounts / creators: [NUMBERS] - Total views to date: [TOTAL] - Views last 2 weeks vs the 2 weeks before: [COMPARE] - Top performing video (views, engagement, age): [DETAILS] - Distinct angles currently running: [NUMBER] - Oldest angle currently active (date first posted): [DATE] - Accounts stuck under 500 views per video: [HOW MANY] - Creators replying to first 30 comments within the first hour: [YES/NO/INCONSISTENT] - Conversion data: [DOWNLOADS / INSTALLS / PURCHASES — if available] Generate: 1. Campaign status: HEALTHY / STALLING / BROKEN — one sentence justification 2. The current winning formula (if identifiable) 3. The 3 highest-leverage actions for the next 7 days — in priority order. Not five. Not ten. Three. 4. Any structural risk needing immediate attention: angle saturation, account suppression, creator disengagement, geographic drift 5. One-line recommendation for the creator manager

Six-Point Viral Checklist

Run any planned video through this before briefing a creator. Most videos that plateau miss at least one.

Run this planned video through the six-point viral checklist before I brief my creator. Planned video concept: [DESCRIBE: hook, format, script outline, visual direction] Product: [PRODUCT NAME] Target audience: [AUDIENCE DESCRIPTION] Evaluate all six points. For each: PASS or FAIL with one sentence. 1. HOOK: Do the first 3-5 seconds create enough curiosity that the viewer feels compelled to stay to the end? Is there a clear visual hook, text hook, and audio that are all emotionally consistent? 2. NO DEAD AIR: Has every moment that doesn't move the story forward been removed — or does this plan contain content that would need cutting before filming? 3. PAYOFF AT THE END: Does the resolution land in the final seconds, not the middle? If the payoff comes early, the viewer satisfies their curiosity and scrolls. 4. RELATABILITY: Does it speak directly to a specific niche rather than trying to appeal broadly? Would someone in the target audience feel this was made specifically for them? 5. SHAREABILITY: Would someone immediately think of a specific person to send this to? Content that is deeply relatable to one community becomes something worth sending — it becomes a vehicle for saying "this is exactly you." 6. REPEATABILITY: Can this format be reproduced consistently, or is it a one-off moment that cannot be replicated? A single viral video with no repeatable structure is not a campaign. Final verdict: ready to brief / needs revision on [specify points] / rework the concept.

07 / Pre-Post

Pre-Post Checklist

Run every video through this before it goes live. If any step fails, fix it before posting. This checklist catches the distribution errors that better scripting cannot fix after the fact. Most videos that plateau below 100,000 views fail at least one of these 12 checks.

Account age and warm-up. Is the account at least 5 days old and properly warmed up? Warm-up means minimum 1 hour per day engaging with niche content — not just watching, but liking, commenting, and watching fully. The For You page should be consistently showing niche-relevant content before posting any branded content.

Pre-post engagement. Have you spent 5-10 minutes engaging with niche content on this account today before posting? Do the same after posting. This signals ongoing human activity to the platform and is required before and after every post.

Three-part hook check. Does the first 2 seconds contain a clear visual hook, text hook, and audio that are all emotionally consistent with each other? A strange visual cannot pair with motivational text. All three must work together or the hook does not pass.

Load-bearing product test. If you removed the product from the video, would it still make sense? If yes, the integration is too weak. The product must be the reason the story exists — not something added at the end. A video that works without the product will get views and zero conversions.

Payoff placement. Does the payoff land in the final 20% of the video — not the middle? If the resolution arrives early, the viewer satisfies their curiosity and scrolls. The algorithm records a low completion rate and suppresses distribution.

Dead air removed. Has all dead air been cut? Filler words, awkward pauses, any moment not moving the story forward. The moment a video slows down is the exact moment the viewer scrolls. Cut filler words; cut drawn-out phrasing; cut anything not earning its place.

Caption format. Is the caption around 10 words, written like a human — not a brand? Include a subtle product CTA roughly half the time. Not every post needs one. Do not start the caption with "I" — TikTok's caption algorithm penalises it.

Hashtags. Are there 3-5 hashtags using a mix of popular, niche, and brand terms? Maximum 5. One broad popular tag, two niche-specific tags, one brand tag. No more than 5.

Font and text placement. Is the video using TikTok's classic font with a black outline? No custom fonts, no coloured text, no glowing effects. Caption placement: slightly below centre. Hook text: slightly above centre. Never at the screen edges.

Banned words check. Does the video use any banned or flagged words — vape, sex, OnlyFans, and others? If yes, substitute letters with symbols or rework the line. Test: search the word on TikTok. Zero results means the platform suppresses it and will suppress your video too.

Duplicate footage check. Is any clip in this video reused footage from a previous video on this account? If yes, re-film it. TikTok recognises duplicate footage and suppresses the newer video. Re-film the same shot as a new recording if the footage is needed again.

Multi-account timing. Are you posting from the same device and Wi-Fi as another account within the last hour? If yes, wait. Back-to-back posting on different accounts from the same device and network flags bot activity. Minimum 1 hour gap between posts on different accounts.

✅ Checklist Prompts 2 Prompts

Pre-Post Checklist Runner

Run any video through all 12 checks. Output: READY TO POST or HOLD — FIX [ITEMS] FIRST.

Run this video through the full pre-post checklist before it goes live. For each check: PASS, FAIL, or FLAG (needs human confirmation). Video details: - Account age and warm-up status: [DETAILS] - Pre-post engagement completed today: [YES/NO] - Hook (text and visual): [DESCRIBE] - Format: [TALKING HEAD / REACTION-DEMO / SKIT / CAROUSEL] - Script summary: [BRIEF OUTLINE] - Where the payoff lands: [SECONDS / POSITION IN VIDEO] - Dead air or filler moments: [YES/NO — describe if yes] - Caption text: [PASTE CAPTION] - Hashtags: [LIST THEM] - Font used: [CLASSIC FONT / OTHER] - Any potentially flagged words: [LIST OR "NONE"] - Any reused footage from previous posts: [YES/NO] - Last post on a different account from this device: [TIME — e.g. "45 minutes ago"] Run all 12 checks. Flag any fails immediately. Final output: READY TO POST or HOLD — FIX [LIST ITEMS] FIRST. If the caption or hashtags fail: generate a replacement caption and hashtag set that passes.

Caption and Hashtag Generator

10-word caption in human tone, correct hashtag mix. Caption rules from Layer 2 built in.

Write a caption and hashtag set for this TikTok video. Caption rules: approximately 10 words. Written like a human — not a brand. Include a subtle product CTA roughly half the time. Do not start with "I." Write 3 options and indicate which include a CTA and which do not. Hashtag rules: 3-5 maximum. Mix: 1 broad popular hashtag (high volume, directly relevant to the niche), 2 niche-specific hashtags (medium volume, directly relevant to the content), 1 brand or product hashtag. Never more than 5. Video content: - What the video is about: [1-2 SENTENCE SUMMARY] - Product featured: [PRODUCT NAME] - Target audience: [DESCRIBE] - Tone of the video: [CASUAL / EDUCATIONAL / SHOCK / STORY / SKIT] Generate: 1. 3 caption options — label which include a CTA and which do not 2. 1 hashtag set of 3-5 tags balancing popular, niche, and brand 3. One warning if any caption option risks triggering a flag

Reference

Exact Benchmarks

Every threshold and benchmark used in the diagnostic prompts above, sourced directly from the Plutus campaign knowledge base. Use these to interpret Claude's outputs and evaluate your own data without guessing.

TikTok Distribution Thresholds

01Under 500 Views

Hook failure. TikTok pushed to its minimum test pool and the video did not earn a second push. Fix: rework the first 1-2 seconds. Different visual, different text, different angle. Do not change the body.

02500 – 3,000 Views

Retention issue. Hook worked but drop-off killed momentum. Video is too long, payoff arrived too early, or there is dead air in the middle. Cut shorter and front-load the payoff.

033,000 – 15,000 Views

Healthy range for a new account or a new angle. Do not change the structure. Evaluate engagement rate before deciding anything else.

0415,000+ Views

Breakout territory. Identify the specific variable that was different and replicate immediately across every account in the program. Move before competitors copy the format.

If most of the last 6 posts are stuck under 300 views regardless of content quality: account problem, not content problem. Fix the account before touching the script.

Metric	Threshold	Interpretation
Engagement rate	Under 2%	No emotional trigger. Add one provocation to the next iteration — not three. One.
Engagement rate	2-5%	Solid. If skewed to likes with almost no comments, add a question or incomplete thought at the end.
Engagement rate	5%+	Strong resonance. If views are low despite high engagement, fix the hook — not the content.
Engagement rate (Spark Ads)	Under 3% at any volume	Weak signal. Do not boost.
Engagement rate (Spark Ads)	7-10%+ at 5k-50k views	Strong Spark Ad signal. Check comment quality and share rate.
Engagement rate (Spark Ads)	5%+ at 50k+ views	Excellent at scale. Test as Spark Ad immediately.
Share rate	Above 1%	Strong shareability. High share rate + high save rate = test as ad immediately.
Save rate	Above 2%	Exceptional content value signal.
30-day campaign benchmark	3 videos with 7%+ like-to-view AND 100k+ views	Campaign on track.
Conversion diagnosis threshold	3,000+ views AND 2%+ engagement	Minimum data required before diagnosing conversion. Below this there is not enough signal.
Angle saturation flag	4+ weeks running, declining views	Not a content quality problem. Angle age problem. Refresh the hook concept first.

Account Rule	Value
Warm-up duration	Minimum 5 days before posting any branded content
Daily warm-up engagement	At least 1 hour per day — like, comment, watch content fully, watch some twice
Pre/post-posting engagement	5-10 minutes engaging niche content before and after every post
Multi-account gap (same device + Wi-Fi)	Minimum 1 hour between posts on different accounts
Account creation/deletion limit	Maximum 1 per 24-hour window
Followers needed for link in bio	1,000 (personal and business accounts both)
Maximum hashtags per post	5
Commercial-only audio (unavailable on personal)	~10-15% of trending audio
Audio available on both account types	~80-90% of top-performing audio
Giveaways	Never — permanently damages audience quality and makes future content invisible to buyers
Minimum video age before diagnosing	48 hours

Creator Pay Rate (April 2026)	Amount
Standard monthly base	~$500/month for 5-7 posts/week
Per-video rate (experienced creators)	$30-40/video
Per-video rate (top demand)	Up to $50/video
View bonus — threshold model (preferred)	$100 per 100,000 views or $250 per 250,000 views
View bonus — CPM model	$1 per 1,000 views across all views
Avoid purely CPM with no base	Attracts low-quality creators, burns out high-quality ones
Payment cadence	Weekly or bi-weekly — monthly reduces motivation and increases drop-off
Ideal sourcing profile	2,000-10,000 followers, 1-5 videos with 100k+ views, shows face on camera, US-based
Creator listing platform	Sideshift (~$150/month) — include visual logo, total monthly pay with bonuses, clear earning potential

Video Production Standard	Value
Hook length	First 3-5 seconds
Body clip length	1-3 seconds per clip
On-screen text per clip	5 words or fewer
Caption length	~10 words, human tone, not brand tone
Font	TikTok classic font, black outline — no custom fonts, no coloured text, no glow effects
Caption placement	Slightly below centre
Hook text placement	Slightly above centre
Preferred filming device	iPhone
Framing	Fill the full frame — no black bars
Effects, animations, filters	Skip entirely for UGC
Editor	TikTok native editor or CapCut
Product reveal timing	Never before the 15-second mark — naming it early gives viewers a reason to leave TikTok

38 Claude Prompts That Replaced Our $10K/Month Production Workflow

What Changed. What Didn't. What This Is.

What's Inside