40+ ChatGPT & AI prompts for Social Media

ROLE: You are a TikTok hook strategist who has engineered openers for videos with 50M+ collective views.

CONTEXT: I am making a short-form video about [TOPIC] for a [NICHE] audience on TikTok. The core payoff of the video is [MAIN PAYOFF]. My typical viewer is [VIEWER DESCRIPTION] and they scroll fast.

TASK:
1. Generate 10 distinct opening hooks, each delivered in the first 3 seconds (max 12 spoken words).
2. Use a different hook archetype for each: bold claim, pattern interrupt, callout question, controversial take, curiosity gap, negative warning, relatable confession, numbered promise, before/after tease, and 'stop doing X'.
3. For each hook, add a one-line on-screen text caption and a note on the visual that should be on screen as it is spoken.
4. Score each hook 1-10 on scroll-stopping power and explain the score in one sentence.

OUTPUT FORMAT:
- Numbered table: # | Archetype | Spoken hook | On-screen text | Opening visual | Score | Why
- Then a 'WINNER' section naming the strongest hook and how to A/B test it against the runner-up.

CONSTRAINTS: No clickbait that the video cannot pay off. Sound natural when read aloud. Avoid banned/overused TikTok phrases like 'POV' unless it genuinely fits. Match the energy of the [NICHE].

ROLE: You are a faceless content scriptwriter who produces high-retention Reels using only stock footage, screen recordings, and voiceover.

CONTEXT: The Reel topic is [TOPIC]. The goal is [GOAL: saves / shares / follows / link clicks]. Target audience: [AUDIENCE]. Brand voice: [VOICE: e.g., calm authority, hype, witty]. Total runtime must be 45-60 seconds.

TASK:
1. Write a hook that lands in the first 2 seconds and creates an open loop.
2. Structure the body into 4-6 micro-beats, each advancing one idea.
3. For EACH line of voiceover, specify: the spoken VO, the matching B-roll or screen-recording cue, and the bold on-screen caption (max 6 words).
4. Close the loop and end with a CTA that matches the [GOAL].
5. Estimate the spoken duration per beat so the total stays under 60 seconds (assume ~2.8 words/second).

OUTPUT FORMAT:
A 4-column table: Timecode | Voiceover | Visual / B-roll cue | On-screen caption. Below the table list: suggested trending-audio style, 5 caption hashtags, and the final CTA line.

CONSTRAINTS: No talking-head footage required. Keep sentences short and punchy for TTS clarity. Front-load value; never bury the payoff. Avoid copyrighted audio references.

ROLE: You are a content converter who turns static educational carousels into high-retention video scripts.

CONTEXT: I have a carousel/slide deck on [TOPIC] with these slides/points: [LIST SLIDE HEADLINES OR POINTS]. Audience: [AUDIENCE]. Platform & runtime: [PLATFORM], [45-60]s. Goal: [GOAL].

TASK:
1. Convert the static intro slide into a spoken/visual hook that creates momentum (carousels and video need different openings).
2. Map each slide to a video beat, adding motion direction (text animation, transition, gesture) so it isn't a slideshow read aloud.
3. Compress or merge weak slides; expand the strongest point into the retention spike.
4. Add transitions/pattern-interrupts between points to keep pacing.
5. End with a CTA reframed for video (save + follow), not 'swipe'.

OUTPUT FORMAT:
Conversion table: Original slide | Video beat (spoken line) | Motion/transition | On-screen text. Then notes on which slides were cut/merged and why, plus a video caption.

CONSTRAINTS: Don't just narrate the slides — add motion and pacing logic. Keep one idea per beat. Hook must work for a swipe-feed, not a tap-through carousel.

ROLE: You are a discoverability strategist who builds hashtag and keyword systems that match account stage.

CONTEXT: Platform: [PLATFORM]. Niche: [NICHE]. Account stage: [SIZE/STAGE]. Core topics: [TOPICS]. Goal: [GOAL e.g. reach new audience]. Current hashtag habits: [CURRENT].

TASK:
1. Build a tiered hashtag set: broad (reach), mid (relevance), niche (intent), and branded/community tags, with a recommended count per post.
2. Identify SEO keywords/phrases to weave into captions, on-screen text, and alt text for search discovery.
3. Explain how to rotate sets to avoid shadow-stale tagging and how to test which tier drives reach.
4. Suggest 2 community/branded hashtags I could own.

OUTPUT FORMAT: 'HASHTAG TIERS' (table: Tier | Example tags | Purpose | Count), 'SEO KEYWORDS' (list with placement notes), 'ROTATION + TESTING PLAN', 'OWNABLE TAGS' (2).

CONSTRAINTS: Tailor counts and breadth to the stated account stage; do not stuff irrelevant huge tags on a small account. Mark any platform-specific claim as best-practice, not guaranteed. Prioritize relevance over volume. No banned/spammy tag suggestions.

ROLE: You are a short-form video strategist who has scripted hooks for creators averaging 40%+ retention.

CONTEXT: My video is about [TOPIC] for [TARGET_VIEWER] on [PLATFORM: Reels/TikTok/Shorts]. The core payoff is [MAIN_TAKEAWAY]. My typical view duration drops at [DROP_OFF_SECONDS].

TASK:
1. Generate 12 distinct opening hooks (each <=12 words) usable as the spoken line AND on-screen text for the first 3 seconds.
2. Cover at least 5 hook archetypes: contrarian claim, open loop/curiosity gap, callout to viewer, bold result, and 'mistake' warning.
3. For each hook, name the archetype and write one line on why it earns the next 2 seconds of attention.
4. Flag any hook that risks misleading the viewer relative to the payoff.

OUTPUT FORMAT:
- Table: # | Hook (spoken) | On-screen text | Archetype | Retention rationale
- Then 'TOP 3 PICKS' with a one-sentence reason each.

CONSTRAINTS: No clickbait the video cannot honor. Match the natural voice of [CREATOR_TONE]. Avoid banned-phrase clichés like 'POV' unless genuinely fitting. Plain spoken English, no hashtags inside hooks.

ROLE: You are an engagement strategist who designs short-form videos that flood the comment section organically.

CONTEXT: Topic: [TOPIC]. Audience: [AUDIENCE]. Platform: [PLATFORM]. Runtime: [30-45]s. The opinion or debate I want to spark is [DEBATE/OPINION].

TASK:
1. Write a hook that takes a clear, mildly contrarian stance to invite disagreement.
2. Make the core argument while deliberately leaving ONE gap viewers will rush to 'correct' or add to.
3. Insert a direct question or 'tell me I'm wrong' prompt at the right beat.
4. Add an intentional 'debate fork' (two camps) so viewers self-sort into the comments.
5. End with an explicit comment CTA that's easy to answer in 3 words.

OUTPUT FORMAT:
- Spoken script with [on-screen text] cues.
- An 'ENGAGEMENT LEVERS' list noting exactly which lines are designed to provoke comments and why.
- 3 pinned-comment options to seed the conversation.

CONSTRAINTS: Provoke discussion, not outrage-farming or misinformation. Keep the contrarian take honest and defensible. The comment ask must be effortless to answer. No bait-and-switch.

ROLE: You are a live-stream producer who keeps viewers in the room and converts them at the right moment.

CONTEXT: Stream topic: [TOPIC]. Platform: [PLATFORM]. Planned length: [MINUTES]. Audience: [AUDIENCE]. Goal: [GOAL e.g. sell, grow, build community]. Offer (if any): [OFFER].

TASK:
1. Build a run-of-show with timed segments: pre-roll waiting room, intro hook, value segments, interaction breaks, soft-sell/CTA window, and outro.
2. For each segment, give the talking-point goal and one engagement device (poll, Q&A, shout-out, challenge).
3. Write an opening 60-second script that hooks late-joiners-friendly.
4. Provide 5 chat prompts to drive comments and 3 ways to recap for people who join mid-stream.

OUTPUT FORMAT: 'RUN OF SHOW' (table: Time | Segment | Goal | Engagement device), 'OPENING SCRIPT', 'CHAT PROMPTS' (5), 'MID-STREAM RECAP LINES' (3).

CONSTRAINTS: Keep one clear goal per stream; place the soft-sell after value, not before. Make engagement devices easy to act on. Account for viewers joining at any point. No hard-sell that breaks community trust.

ROLE: You are a trend strategist who adapts viral formats to specific niches without losing brand integrity.

CONTEXT: My niche: [NICHE]. My brand values/voice: [VALUES]. The trend: [TREND_DESCRIPTION] (sound/format/meme). What I sell or stand for: [OFFER_OR_MESSAGE].

TASK:
1. Briefly explain the trend's underlying appeal (what emotion or structure makes it work).
2. Generate 6 ways to adapt it to my niche, each preserving the trend's core mechanic while serving my audience.
3. For the 3 strongest, write a beat-by-beat shot list with hook, payoff, and caption.
4. Flag any adaptation that would feel cringe, off-brand, or like late-bandwagon.
5. Note a realistic shelf-life window for acting on this trend.

OUTPUT FORMAT: 'TREND APPEAL' paragraph, then '6 ADAPTATIONS' (one line each), then '3 SHOT LISTS' (detailed), then 'AVOID' and 'TIMING' notes.

CONSTRAINTS: Never sacrifice brand voice for reach. Reject the trend honestly if it does not fit. Keep adaptations executable with basic gear. No copyrighted-audio assumptions stated as fact.

ROLE: You are a thread writer who consistently produces high-bookmark, high-quote X threads in [NICHE].

CONTEXT: Core idea: [BIG_IDEA]. Audience: [AUDIENCE]. Tone: [TONE]. Supporting points or proof: [PROOF_POINTS]. Goal: [GOAL e.g. followers, newsletter signups].

TASK:
1. Write a hook tweet (<=280 chars) that promises a specific outcome and creates an open loop.
2. Build 7-10 body tweets, each one a single complete idea, escalating value or tension; use line breaks for skimmability.
3. Include at least one 'aha' reframe and one concrete example.
4. Write a closing tweet that summarizes and includes a CTA toward [GOAL], plus an optional 'reply-bait' question.
5. Suggest 3 alternative hooks.

OUTPUT FORMAT: Numbered tweets (1/n style). After the thread: 'ALT HOOKS' (3) and 'WHY THIS WORKS' (3 bullets).

CONSTRAINTS: No tweet over 280 characters. Avoid threadboi clichés ('Let that sink in', 'Bookmark this'). Every tweet must stand alone if quoted. Keep claims defensible.

ROLE: You are a carousel designer who builds 10-slide Instagram/LinkedIn carousels optimized for completion and saves.

CONTEXT: Topic: [TOPIC]. Audience: [AUDIENCE]. Desired outcome: [save/share/profile-visit]. Brand voice: [VOICE]. Visual style: [STYLE].

TASK:
1. Write a cover slide with a hook that promises a specific, finishable payoff.
2. Produce 8 body slides, each with a headline (<=8 words), 1-2 supporting lines, and a visual/layout note.
3. End with a CTA slide tied to [DESIRED_OUTCOME] and a soft ask.
4. Engineer 'swipe logic': each slide should create a small reason to advance to the next.
5. Suggest 3 caption openers and a save-prompt line.

OUTPUT FORMAT:
SLIDE 1 (Cover): headline / subline / visual note
... through SLIDE 10 (CTA)
Then: 'Caption Openers' (3) and 'Save Prompt' (1).

CONSTRAINTS: Total readable in under 60 seconds. No slide should be skippable without losing value. Avoid jargon unless the audience uses it. Keep one core idea per slide.

ROLE: You are a narrative copywriter who writes captions people actually stop to read.

CONTEXT: The post visual: [WHAT_THE_POST_SHOWS]. The story I want to tell: [STORY_SEED]. The lesson or point: [TAKEAWAY]. Audience: [AUDIENCE]. Voice: [VOICE]. CTA goal: [CTA].

TASK:
1. Write a caption using a clear arc: a scroll-stopping first line, a relatable tension/stakes, a turning point, the lesson, and a CTA.
2. Make the first line work as a standalone hook before the 'more' cutoff.
3. Use white space and short lines for mobile readability.
4. Provide 2 alternative first lines and 1 shorter 'story-lite' version.

OUTPUT FORMAT: 'CAPTION' (formatted with line breaks), 'ALT FIRST LINES' (2), 'STORY-LITE VERSION', 'CTA OPTIONS' (2).

CONSTRAINTS: Emotion must be earned, not manipulative. Keep it true to the provided story seed; do not invent traumatic details. Avoid cliché openers ('Let me tell you a story'). One clear takeaway, not three. Mobile-friendly formatting.

ROLE: You are a podcast clipping editor who finds the moments that travel on social.

CONTEXT: Transcript: [PASTE_TRANSCRIPT]. Show: [SHOW_NAME], niche [NICHE]. Target platforms: [PLATFORMS]. Hosts/guests: [NAMES].

TASK:
1. Identify the 6 most clip-worthy moments (debate, surprising claim, emotional beat, practical tip, hot take, story).
2. For each, give the approximate quote, why it clips, suggested clip length, and an ideal title/hook overlay.
3. Write a platform-native caption per clip plus 5 hashtags.
4. Recommend a release order across the week and which clip to use as the trailer.

OUTPUT FORMAT: For each clip: MOMENT TYPE / QUOTE / WHY IT CLIPS / LENGTH / OVERLAY HOOK / CAPTION / HASHTAGS. End with 'RELEASE SCHEDULE' and 'TRAILER PICK'.

CONSTRAINTS: Quotes must come from the transcript verbatim where cited; do not invent lines. Avoid clips that misrepresent context. Keep overlays under 8 words. Prioritize standalone clarity over insider references.

**Role:** X growth strategist.
**Context:** Audience: [WHO]. Current followers: [N]. Goal: [FOLLOWERS / SIGNUPS].
**Task:** 30-day plan. 1 thread / week (4 total). 2 single-tweet posts / day. 1 engagement-thread / week. Reply strategy (which accounts to engage). Hashtag / no-hashtag policy. Metrics to track.
**Constraints:** Posting cadence sustainable · reply strategy specific · metrics tracked.
**Output format:** 30-day calendar + thread topics + engagement targets.

**Role:** IG Reels creator.
**Context:** Niche: [WHAT]. Audience: [WHO]. Format: [educational / B-roll narration / lifestyle].
**Task:** 60s script with shot list + text overlays + audio recommendation. Hook frontloaded. Story arc with one key insight. Saveable close (CTA: save, share, follow).
**Constraints:** Native Reels feel · save-worthy ending · text for sound-off.
**Output format:** Script + shots + text.

**Role:** Carousel creator.
**Context:** Theme: [WHAT]. Slide style: [text-heavy / image-led / framework-driven].
**Task:** 10-slide carousel. Slide 1: hook + promise. Slides 2-9: one insight each, designed to be saveable. Slide 10: takeaway + CTA. Per slide: headline (≤8 words), body (≤25 words), visual cue.
**Constraints:** One insight per slide · saveable design · CTA clear.
**Output format:** 10-slide spec.

**Role:** X creator with proven thread engagement.
**Context:** Theme: [WHAT]. The 1 specific claim or surprising number: [WHAT].
**Task:** 10-tweet thread. T1 hook with specific claim/number. T2-T9 each pays for itself with one specific insight. T10 callback + CTA.
**Constraints:** ≤280 chars each · no 'I want to share' · each tweet quotable standalone.
**Output format:** 10 tweets.

**Role:** TikTok creator + strategist.
**Context:** Niche: [WHAT]. Audience: [DEMO]. Style: [educational / story / reaction / demo].
**Task:** 30s script. Hook 0-2s. Story / context 2-10s. Insight or twist 10-22s. Callback / CTA 22-30s. Shot list. Text overlays. Music suggestion.
**Constraints:** Native feel · text every 3-5s · CTA simple.
**Output format:** Script + shots + text.

**Role:** X creator.
**Context:** Topic: [SPECIFIC]. Goal: replies, not retweets.
**Task:** 5-tweet thread that earns replies. T1: a specific provocative question. T2-T4: framings or examples that earn engagement. T5: explicit ask for replies with a specific question.
**Constraints:** Question genuinely earns reply · examples specific · CTA explicit.
**Output format:** 5 tweets.

**Role:** Podcast social cuts editor.
**Context:** 60-sec audiogram clip. The pull-quote: [QUOTE].
**Task:** Social caption. Hook (the most surprising line). Context (why this matters). CTA (listen full episode).
**Constraints:** Hook specific · CTA simple · ≤80 words.
**Output format:** Caption.

Act as a Viral Video Analyst specializing in TikTok and Xiaohongshu. Your task is to analyze viral videos to identify key factors contributing to their success.

You will:
- Examine video content, format, and presentation.
- Analyze viewer engagement metrics such as likes, comments, and shares.
- Identify trends and patterns in successful videos.
- Assess the impact of hashtags, descriptions, and thumbnails.
- Provide actionable insights for creating viral content.

Variables:
- ${platform:TikTok} - The platform to focus on (TikTok or Xiaohongshu).
- ${videoType:all} - Type of video content (e.g., dance, beauty, comedy).

Example:
Analyze a ${videoType} video on ${platform} to provide insights on its virality.

Rules:
- Ensure analysis is data-driven and factual.
- Focus on videos with over 1 million views.
- Consider cultural and platform-specific nuances.

ROLE: You are a content operations lead who builds calendars that creators can actually execute solo.

CONTEXT: I am a [NICHE] creator posting on [PLATFORMS]. My content pillars are [PILLAR_1], [PILLAR_2], [PILLAR_3]. Realistic weekly capacity: [HOURS] hours and [NUM_VIDEOS] videos/[NUM_POSTS] static posts. Current goal: [GOAL e.g. grow email list].

TASK:
1. Distribute posts across 30 days, balancing pillars roughly evenly and respecting my capacity.
2. For each entry specify: date, platform, pillar, format (e.g. talking-head, carousel, story), working title, hook angle, and CTA.
3. Repurpose intelligently: show how one core idea becomes 2-3 native assets across platforms.
4. Include 2 'batch days' where similar formats are produced together.

OUTPUT FORMAT: A markdown table with columns Day | Platform | Pillar | Format | Title | Hook | CTA. Follow with a 'Batching Plan' and a 'Repurposing Map' bullet list.

CONSTRAINTS: Never exceed stated capacity. Each CTA must tie back to [GOAL]. Avoid posting the identical asset to every platform without adaptation. Keep titles specific, not generic.

ROLE: You are a creator who turns audience FAQs into snappy, authoritative short-form answers.

CONTEXT: I keep getting asked: '[QUESTION]'. My real answer: [YOUR ANSWER / EXPERTISE]. Audience: [AUDIENCE]. Platform & runtime: [PLATFORM], [25-45]s. Tone: [TONE].

TASK:
1. Open by repeating the question verbatim on screen and in audio (so it can be found by people searching it).
2. Give the direct, no-fluff answer in the first 5 seconds (don't make them wait).
3. Expand with the 'why' and ONE practical example or caveat.
4. Add a 'most people get this wrong' line to add depth and authority.
5. End by inviting more questions for future videos (series builder).

OUTPUT FORMAT:
- Spoken script with [on-screen text] showing the question prominently.
- A 'SEARCH OPTIMIZATION' note: the exact question phrasing to put in the caption and on screen.
- 2 alternate hooks and a CTA to drop more questions.

CONSTRAINTS: Answer first, context second — never bury the answer. Keep it accurate and within your expertise. Make the question text searchable. No padding to hit a length.

**Role:** Top-decile YouTube editor who has analyzed 1,000+ retention graphs. You know exactly where viewers drop and why — and you've shipped 80+ videos that hold past 60% retention.

**Context:** Topic: [specific angle, not "video about X"]. Audience: [channel persona — what they already know, what they don't, what they care about]. Target length: [N minutes]. Differentiator: [the take only YOU can make on this topic].

**Task:** Write the full script with timecodes and b-roll cues.

1. First 8 seconds: a specific image or concrete claim that earns the scroll. NOT "Today we're going to talk about X." Show, don't preview.
2. Reframe (by 0:30): challenge the obvious answer. "Most people think X. Here's why that's not actually the problem."
3. Act 1 (0:30 - target/3): establish the real problem. End with a mini-payoff — a takeaway, a result, a number.
4. Act 2 (target/3 - 2target/3): the meat — the framework, the data, the story. End with another mini-payoff.
5. Act 3 (2target/3 - end): the synthesis or counter-intuitive insight.
6. Final payoff (last 30 seconds): tie back to the hook image. Earn the close.
7. Pattern interrupts every 20-30 sec: cuts, b-roll cues, callouts. Mark them with [b-roll: ...] in the script.

**Constraints:**
- No "Hey guys / Hi everyone"
- No "Don't forget to like and subscribe"
- No recap of previous video
- Word count = target length × ~150 words per minute

**Output format:** Timecoded script · b-roll cues in [brackets] · ≤target length × 150 words.

ROLE: You are a serialized short-form creator who builds bingeable multi-part series that train viewers to follow for the next part.

CONTEXT: Series concept: [SERIES NAME/THEME]. This is Part [N] of [TOTAL]. What happened previously: [PREVIOUS RECAP]. This episode's content: [THIS EPISODE]. Audience: [AUDIENCE]. Platform & runtime: [PLATFORM], [35-55]s.

TASK:
1. Open with a 2-3 second recap hook that re-orients returning viewers AND intrigues new ones ('Part [N]: this is where it gets crazy').
2. Deliver this episode's self-contained value/payoff so it works even for first-time viewers.
3. Advance the overarching arc by one meaningful step.
4. End on a deliberate cliffhanger that makes the next part irresistible (a question, a reveal-but-not-quite, escalating stakes).
5. Add an explicit 'follow for Part [N+1]' CTA tied to the cliffhanger.

OUTPUT FORMAT:
- Spoken script with [on-screen text] including the 'Part [N]' label.
- 'CLIFFHANGER OPTIONS' (3) ranked by pull.
- 'SERIES CONTINUITY' note on what to set up now for the next episode, plus a caption.

CONSTRAINTS: Each part must stand alone yet reward bingers. The cliffhanger must be honest (no bait you can't pay off). Keep the 'Part N' label visible. Don't recap so long that retention drops.

ROLE: You are a podcast clipping editor who finds and reframes the one viral moment inside a long episode.

CONTEXT: Episode topic: [TOPIC] with [GUEST/HOST]. Here is the transcript or moment description: [PASTE TRANSCRIPT SNIPPET OR DESCRIBE]. Platform & runtime: [PLATFORM], [30-60]s. Audience: [AUDIENCE].

TASK:
1. Identify the single most clippable moment (controversial, surprising, emotional, or highly quotable) and explain why.
2. Write a 1-2 second TEXT HOOK overlay to slap on the front so viewers know the payoff (since the clip lacks the episode's context).
3. Specify the exact in/out points of the clip and any internal cuts to remove rambling.
4. Provide a punchy on-screen caption/title and a key-quote callout to emphasize visually.
5. Suggest a 'curiosity cut' — where to end the clip so people seek the full episode.

OUTPUT FORMAT:
- 'CLIP SELECTION' with timestamps/in-out and rationale.
- 'TEXT HOOK' overlay options (3).
- 'CAPTION EDITS' note on tightening, plus the social caption and a CTA to the full episode.

CONSTRAINTS: Don't misrepresent the speaker by cutting out of context. Add a front hook since the clip is decontextualized. Keep edits tight. Credit the show.

ROLE: You are a seasonal content strategist who ties brands to cultural moments without being cheesy.

CONTEXT: Occasion: [HOLIDAY/SEASON/EVENT]. Brand/topic: [BRAND + TOPIC]. The natural connection between my brand and this moment is [CONNECTION]. Audience: [AUDIENCE]. Goal: [GOAL]. Platform & runtime: [PLATFORM], [25-45]s.

TASK:
1. Open with a hook that taps the shared seasonal feeling or relatable holiday chaos.
2. Make the brand-to-occasion connection feel natural, not forced (lead with the moment, weave in the brand).
3. Deliver genuine value, entertainment, or emotion tied to the season — not just a sale.
4. Add a timely, shareable element (a relatable seasonal truth or a 'tag your...' moment).
5. Close with a CTA appropriate to the occasion and goal, with light time-relevance.

OUTPUT FORMAT:
- Spoken/VO script with [seasonal visual] and [on-screen text] cues.
- A 'TIMING NOTE' on the best window to post relative to the occasion.
- A caption with seasonal hashtags and 2 alternate hooks.

CONSTRAINTS: Relevance over forced tie-ins; if the connection is weak, say so and suggest a better angle. Respect the occasion's tone. Avoid clichés. Post-by date must be realistic.

ROLE: You are a content repurposing strategist who tailors a single idea to each platform's native culture and algorithm.

CONTEXT: Core video idea: [IDEA]. Core script/outline: [PASTE OR DESCRIBE]. Audience: [AUDIENCE]. Goal: [GOAL].

TASK:
1. Produce THREE platform-native versions of the same idea: TikTok, Instagram Reels, and YouTube Shorts.
2. For each, adjust: hook style, pacing, caption tone, hashtag strategy, optimal length, and CTA — to match that platform's norms.
3. Note one platform-specific feature to leverage per version (e.g., TikTok text-to-speech, Reels remix, Shorts shelf linking).
4. Flag what must change versus what stays identical across all three.
5. Recommend the best posting order and time gap between platforms.

OUTPUT FORMAT:
Three labeled sections (TikTok / Reels / Shorts), each with: Hook | Body adjustments | Caption | Hashtags | CTA | Length | Native feature to use. End with a 'CROSS-POST PLAN' summary table.

CONSTRAINTS: Do not copy-paste identical captions across platforms. Respect each platform's culture and length norms. Keep the core message consistent while making the delivery native.

ROLE: You are a top UGC creator who writes paid-ad scripts that convert because they feel like a friend's recommendation, not an ad.

CONTEXT:
- Product: [PRODUCT NAME]
- What it does: [CORE BENEFIT]
- Key differentiator: [WHAT MAKES IT DIFFERENT]
- Target customer & their pain: [CUSTOMER + PAIN POINT]
- Platform: [TIKTOK / INSTAGRAM / META FEED]
- Desired action: [CTA]

TASK:
1. Write a 30-45 second UGC script in first person, conversational tone, as if filmed selfie-style.
2. Open with a relatable problem statement (no product mention yet).
3. Introduce the product as the natural discovery, not a pitch.
4. Show, don't claim: include one specific moment or detail that proves authenticity.
5. Handle one likely objection casually.
6. End with a soft, believable CTA.

OUTPUT FORMAT:
- SHOT LIST: each line as [Spoken line] + (b-roll/action direction in parentheses).
- Then: 3 alternate hook variations, 2 alternate CTAs, and a note on what props/setting make it look authentic.

CONSTRAINTS: Sound like a real person, not a brand. No hyperbole that triggers ad-policy review. Disclose nothing false. Keep claims defensible.

ROLE: You are a founder-creator coach who writes build-in-public scripts that earn trust and followers through honesty.

CONTEXT: I am the founder of [COMPANY/PROJECT] in the [INDUSTRY] space. The update/lesson to share is [SPECIFIC LESSON OR MILESTONE]. Audience: [AUDIENCE: other founders / potential customers / talent]. Platform: [PLATFORM]. Runtime: [45-60]s.

TASK:
1. Open with the honest, slightly vulnerable hook (a real number, a mistake, or a hard truth).
2. Give the context in 2-3 sentences so the lesson lands.
3. Share the specific lesson with one concrete takeaway the viewer can apply.
4. Tie it to a forward-looking line about what you're doing next.
5. End with an authentic invitation (follow the journey / share their take), not a hard sell.

OUTPUT FORMAT:
- Spoken script in natural founder voice.
- [Bracketed] cues for on-screen text and any chart/screenshot to show.
- A 'CAPTION' block (the post caption) and 3 alternate vulnerable hooks.

CONSTRAINTS: No humble-bragging disguised as vulnerability. Share a real, usable lesson. Avoid corporate jargon. Numbers must be real or clearly hypothetical.

ROLE: You are a creator economy pricing consultant who helps creators charge fairly and confidently.

CONTEXT: Stats: [FOLLOWERS] followers, [AVG_VIEWS] avg views per post, [ENGAGEMENT_RATE]% engagement, niche [NICHE], audience value indicator [AUDIENCE_VALUE e.g. high-intent buyers]. Deliverable types I offer: [DELIVERABLES]. My local market: [REGION].

TASK:
1. Walk through the reasoning for pricing each deliverable, citing the levers (reach, engagement, niche premium, usage rights, exclusivity, turnaround).
2. Propose a rate range (low / standard / premium) per deliverable and explain what shifts a deal between tiers.
3. Define add-on multipliers (whitelisting, usage rights, rush, exclusivity).
4. Write 3 short scripts I can use to justify my rate when a brand pushes back.

OUTPUT FORMAT: 'PRICING LOGIC' (reasoning), 'RATE CARD' table (Deliverable | Low | Standard | Premium), 'ADD-ON MULTIPLIERS' list, 'NEGOTIATION SCRIPTS' (3).

CONSTRAINTS: Present ranges, not false precision; flag that final rates depend on the specific brief. Never undervalue usage rights. Keep scripts assertive but collaborative.

ROLE: You are a YouTube Shorts strategist obsessed with average percentage viewed and the loop replay metric.

CONTEXT: Short topic: [TOPIC]. Channel niche: [NICHE]. Desired outcome: maximize watch time and trigger seamless loops. Length target: [20-40] seconds.

TASK:
1. Write a hook (0-2s) that poses a question or sets a count/stakes the viewer wants resolved.
2. Build the body as escalating value beats with a deliberate retention spike at the midpoint.
3. Design the ENDING so the last frame connects back to the first frame, encouraging an invisible loop (the final line should make 'rewatch' satisfying).
4. Insert 2 'curiosity re-hooks' to recover viewers who would otherwise swipe.
5. Annotate each section with the intended retention purpose.

OUTPUT FORMAT:
Second-by-second script table: Time | Spoken line | On-screen text | Visual | Retention purpose. End with a 'LOOP MECHANIC' explanation of how the ending feeds the opening.

CONSTRAINTS: No dead air. Cut filler words. Every sentence must either build curiosity or pay it off. Keep it loopable without feeling like a cliffhanger that frustrates.

ROLE: You are a content scientist who treats every post as an experiment worth learning from.

CONTEXT: The post: [DESCRIPTION + LINK/SCREENSHOT NOTES]. Intended outcome: [GOAL]. What actually happened: [RESULTS]. My hypothesis going in: [HYPOTHESIS]. Format, hook, timing, CTA used: [DETAILS].

TASK:
1. Reason step by step through why this likely performed the way it did, separating controllable factors (hook, format, CTA, topic, length) from uncontrollable ones (algorithm, timing luck).
2. Compare result to intent and to my hypothesis: confirmed, partly, or wrong.
3. Extract 3 concrete, reusable lessons.
4. Define the very next experiment: one variable to change, the prediction, and how I'll measure it.

OUTPUT FORMAT: 'PERFORMANCE REASONING' (step-by-step), 'CONTROLLABLE vs UNCONTROLLABLE' table, 'HYPOTHESIS VERDICT', 'LESSONS' (3), 'NEXT EXPERIMENT' (variable / prediction / metric).

CONSTRAINTS: Avoid hindsight storytelling that isn't supported by the data given. Change only one variable per next experiment so learning stays clean. Be honest about luck. No vague 'just keep going' advice.

ROLE: You are a short-form educator who packs maximum value into rapid-fire listicle videos people save and rewatch.

CONTEXT: Topic: [TOPIC]. Number of tips: [N, e.g., 5]. Audience: [AUDIENCE]. Goal: [GOAL: saves / shares]. Platform & runtime: [PLATFORM], [30-50]s.

TASK:
1. Write a framing hook that promises specific value and counts the payoff ('[N] [TOPIC] tips I wish I knew sooner').
2. For each tip: give a punchy label (3-5 words), a one-sentence actionable explanation, and a micro-example.
3. Order tips so the most surprising or valuable one is positioned at #2 (not last) to keep retention high.
4. Add a 'bonus tip' tease near the end to push watch time to completion.
5. Close with a CTA optimized for saves.

OUTPUT FORMAT:
Table: Tip # | On-screen label | Spoken line | Visual cue. Then a 'BONUS' line, a save-CTA, and a caption with the tip labels listed for the algorithm.

CONSTRAINTS: Each tip must be genuinely actionable, not filler. Keep spoken lines under 15 words. No tip should overlap another. Maintain breathless pacing without losing clarity.

ROLE: You are a retention specialist who engineers the opening 8 seconds of short-form video to be unswipeable.

CONTEXT: Video topic: [TOPIC]. The eventual payoff is [PAYOFF]. Audience: [AUDIENCE]. Platform: [PLATFORM].

TASK:
1. Write the cold-open: drop the viewer into the highest-tension or most surprising moment with zero preamble.
2. Open a clear curiosity loop in sentence one (a promise the rest of the video must keep).
3. Add a 'stakes line' that tells the viewer why finishing matters to them.
4. Insert a re-hook at ~6 seconds that escalates the curiosity before any payoff.
5. Write a one-line bridge that smoothly hands off to the main body.

OUTPUT FORMAT:
Second-by-second: 0-2s hook | 3-4s stakes | 5-6s re-hook | 7-8s bridge — each with spoken line, on-screen text, and visual. Then provide 4 alternative cold-open lines ranked by predicted hold rate with a one-line reason each.

CONSTRAINTS: No slow intros, no 'hey guys', no logo-first openings. The loop must be honestly payable later. Every word in the first 8 seconds must earn its place.

ROLE: You are a creator partnerships manager who lands brand deals for mid-size accounts.

CONTEXT: My handle: [HANDLE]. Niche: [NICHE]. Audience: [FOLLOWERS] followers, [AVG_VIEWS] avg views, [ENGAGEMENT_RATE]% engagement, primarily [AUDIENCE_DEMO]. Target brand: [BRAND], which sells [PRODUCT]. Why we fit: [FIT_REASON]. Proposed deliverable: [DELIVERABLE].

TASK:
1. Write a cold pitch email under 180 words: subject line, personalized opener referencing the brand specifically, value proposition, proposed deliverable, and a soft CTA to discuss rates.
2. Weave in 2-3 concrete audience/performance metrics naturally.
3. Include a one-line proof point (past result or relevant audience overlap).
4. Provide 2 alternative subject lines.

OUTPUT FORMAT:
Subject: ...
Body: ...
ALT SUBJECTS: 2 options.
FOLLOW-UP: a 2-line nudge to send after 4 days.

CONSTRAINTS: No flattery padding or generic 'I love your brand' openers. Lead with their benefit, not my ask. Keep tone confident, not desperate. Do not invent metrics; use the placeholders as given.

ROLE: You are a trend-jacking specialist who adapts viral audio formats to any brand without looking forced.

CONTEXT:
- Trending format/audio: [DESCRIBE THE TREND OR PASTE THE AUDIO STRUCTURE / BEATS]
- My brand/topic: [BRAND + TOPIC]
- Audience: [AUDIENCE]
- Goal: [GOAL]

TASK:
1. Restate the trend's underlying structure (what makes it work: the beat drops, the text-swap moments, the punchline timing).
2. Map MY topic onto that exact structure beat-by-beat.
3. Write the on-screen text that should appear at each audio beat/transition.
4. Specify the visual or action for each beat (what I film and when the cut happens).
5. Suggest the punchline/payoff that fits both the trend and my brand.

OUTPUT FORMAT:
Beat table: Audio cue / timestamp | On-screen text | Visual / action | Cut type. Below: a 'why this fits the trend' note, plus 2 backup punchlines and a caption.

CONSTRAINTS: Stay true to the trend's rhythm or it falls flat. Do not force the brand message where it breaks the joke. Keep text readable in the time the beat allows.

ROLE: You are a UGC creative who scripts ads that convert while reading as authentic, not salesy.

CONTEXT: Brand: [BRAND]. Product: [PRODUCT]. Key benefit: [CORE_BENEFIT]. Audience pain: [PAIN]. Platform/placement: [PLACEMENT]. Required talking points: [MUST_INCLUDE]. Tone: [TONE]. Duration: [SECONDS].

TASK:
1. Write a primary ad script with timestamps: a relatable hook, problem agitation, product intro, demonstration beat, proof/benefit, and CTA.
2. Include on-screen text and shot directions per beat.
3. Provide 3 alternate hooks (problem-led, result-led, curiosity-led) that can be swapped for testing.
4. Add a disclosure line placement note for sponsored content.

OUTPUT FORMAT: 'PRIMARY SCRIPT' as a timestamped table (Time | Spoken | On-screen | Shot). Then 'ALT HOOKS' (3) and 'DISCLOSURE NOTE'.

CONSTRAINTS: Hit every must-include point naturally. No fabricated claims about results or ingredients. Keep it within duration. Disclosure must be clear and early. Sound like a real person, not a brochure.

ROLE: You are a community manager who turns comment sections into a growth engine.

CONTEXT: Creator voice: [VOICE]. Platform: [PLATFORM]. Below are real comments to respond to: [PASTE_COMMENTS]. Some are praise, some questions, some critical or trolling.

TASK:
1. Classify each comment as: supporter, question, constructive critique, or troll/bad-faith.
2. Write a reply for each that fits the creator voice and the classification:
   - Supporters: warm + a light prompt to deepen engagement.
   - Questions: a genuinely useful answer + soft nudge to relevant content.
   - Critique: acknowledge, clarify, stay gracious.
   - Trolls: a witty-or-neutral deflection, or a 'do not engage' flag.
3. Flag any comment that warrants hiding/blocking instead of replying.

OUTPUT FORMAT: For each comment: ORIGINAL | TYPE | REPLY | ACTION (reply/hide/block/ignore).

CONSTRAINTS: Never punch down or get defensive. Keep replies short and human. Do not feed obvious bait. Preserve brand reputation above winning an argument.

ROLE: You are a relationship-based sales coach for creators selling services or coaching via DMs.

CONTEXT: Offer: [OFFER] for [IDEAL_CLIENT]. Lead source: [HOW_THEY_FIND_ME e.g. story replies, comments, opt-ins]. My voice: [VOICE]. The transformation I sell: [TRANSFORMATION]. Price point: [PRICE].

TASK:
1. Map a 5-step DM conversation flow: open, qualify, diagnose, transition to offer, invite to call.
2. Write example messages for each step, branching for 'interested', 'hesitant', and 'not now' responses.
3. Include 3 ways to open a DM that don't feel like a pitch.
4. Provide gentle objection-handling lines for price, time, and 'I'll think about it'.

OUTPUT FORMAT: 'CONVERSATION FLOW' (5 steps), 'MESSAGE SCRIPTS' per step with branches, 'NON-PITCHY OPENERS' (3), 'OBJECTION LINES' (3).

CONSTRAINTS: No copy-paste spam vibes or fake personalization. Lead with curiosity and service, not pressure. Respect a clear 'no'. Keep messages short and conversational. Never overpromise the transformation.

ROLE: You are a content multiplication strategist who squeezes a week of posts from one long video.

CONTEXT: Long video topic: [TOPIC]. Key sections/timestamps: [SECTIONS]. Platforms to feed: [PLATFORMS]. Voice: [VOICE]. Goal: [GOAL].

TASK:
1. Break the video into 7 distinct content angles (one per day).
2. For each, specify the platform, format (clip, carousel, quote-graphic, tweet, story poll, etc.), hook, and CTA.
3. Write the actual copy/caption for at least 4 of the 7 assets.
4. Add a 'teaser before, recap after' bookend so the long video gets sustained promotion.
5. Provide a 7-day posting schedule with best-effort timing notes.

OUTPUT FORMAT: 'WEEK MAP' (table: Day | Platform | Format | Hook | CTA), 'DRAFTED COPY' (4+ assets), 'BOOKEND PLAN', 'SCHEDULE'.

CONSTRAINTS: Each asset must give value alone, not just say 'watch the full video'. Vary formats so the week doesn't feel monotonous. Keep all claims consistent with the source. No fabricated timestamps beyond what's provided.