StackVisual
The Visual stack
Image-first reach. Instagram, Pinterest, and Copy, plus Nexin Core.
Founding price $27
Next: $37 per module
47 of 50 left at $27
3 claimed
01What's included
4 modules. One payment.
CORE
FREE
Nexin Core
Interviews you and turns the answers into a brand file every other module reads.
core tool · 23 prompts
MOD/01
$27
Reel scripts, carousels and captions with hooks built for the first second of the feed.
instagram tool · 26 prompts
MOD/03
$27
Search-first pins that keep sending traffic for months, not hours.
pinterest tool · 30 prompts
ADDON
$17
Copy
Storytelling frameworks, proven copy structures and fill-in hook templates that plug into every module.
copy tool · 27 prompts · 248 hook templates
02What you can make
Every job in the Visual stack.
Each job is a method your AI follows in Claude or ChatGPT, reading your brand file first. Ask for the piece you need and it picks the job.
Nexin Core
- Competitor check. Maps about 10 accounts in the niche from what the user pastes (or web search when available), sorts them into twins, reply targets and gaps, tears down their best posts by hook, format, chain and CTA with a take-versus-leave call on each, and names the angle nobody owns.
- Create brand. Builds a brand from zero for a personal brand, a persona-led or faceless account, a business or a product: generates 8 to 10 niche concepts in a comparison table, shortlists 2 or 3, creates names and handles (15 characters or fewer, pronounceable, with collision flags), then writes positioning, an enemy, 3 to 5 pillars written as tensions and visual basics (2 fonts, 3 colors) into nexin-brand.md.
- Create persona. Builds the speaker every Nexin job writes as, for any account: a real person (self), a persona or avatar (alter ego or fictional character), a brand speaking as "we", or a product.
- Interview. Runs the Nexin brand interview one question per message, starting with who speaks (you, a persona, the brand or the product) and how they show up on camera (face, voice or none), then adapting every later question to that speaker.
- Proof ledger. Sorts every claim and story the brand makes or wants to make, for any speaker (a real person, a persona, the brand or the product), into showable (a fact with a source, stated as fact), believed (only written as a named [placeholder] until proven) and forbidden (never written, with a safe rewrite), and writes nexin-proof.md with proof-post shapes and number-free versions of each claim, applying the speaker's story_rights so no first-person story goes beyond what the speaker may tell.
- Publish gate. Pre-publish gate for any Nexin asset (post, carousel, script, pin, article, email, sales page): checks invented proof against nexin-proof.md, income and health claims, first-person stories against the speaker's story_rights, platform limits, link placement, disclosure, AI-writing tells and voice drift against nexin-persona.md (the speaker file, for a real person, persona, brand or product), then returns PASS or a fix list sorted by severity with exact rewrites.
- Validate brand. Scores a brand or niche out of 30 on six criteria (pain, ability to pay, content supply, speaker fit, differentiation, product path), 1 to 5 each, with every assumption stated, and returns a go, fix or rethink verdict with the cheapest test for each weak score.
- Validate persona. Scores nexin-persona.md, the speaker file, for any speaker type (self, persona, brand or product) on four criteria (distinctiveness, consistency, claim safety, sellability), 5 points each for a total out of 20, with claim safety as a hard gate that checks every claim and story against the speaker's story_rights, then rewrites the weak parts and returns an updated persona file with a changelog.
- Voice of customer. Mines pasted comments, reviews, DMs, survey answers and forum threads into nexin-voc.md: verbatim phrases with source and tier (own customers, then proxy products, then forums), sorted into pains, desires, objections, triggers and vocabulary, with frequency counts.
- Winner decoder. Takes 10 to 30 of the user's own posts with metrics, computes the median, flags winners at 2x the median or more, decodes each winner's hook family, structure and format against the rest, and writes nexin-winners.md with do-more and stop rules plus hypotheses to test.
- Caption. Writes Instagram captions in three lengths (short, medium, long-caption sales variant) with the hook inside the first ~125 characters, a body matched to the post format, one CTA from the funnel file, and 3 to 5 niche hashtags or none, with character counts computed.
- Carousel. Builds Instagram carousels of exactly 3, 5, 7 or 10 slides, either educational (save-driven, with a recap keeper slide) or sales (context, tension, value, proof, objection, action), one idea per slide at about 30 words max, 4:5 at 1080x1350, with visual notes and a caption.
- Dm keyword. Builds the comment-keyword to DM flow for Instagram, with the keyword choice, 5 public comment replies, the first DM, delivery, a 24-hour follow-up and a day-3 offer step, in free (freebie) and paid (offer) variants, ready for a manual routine or an automation tool.
- Highlights. Plans 5 to 7 Instagram story highlights (start here, proof, offer, free entry, FAQ, process, topic) with order, names of 10 characters or fewer, cover text and design, the stories that feed each one, a frame-by-frame outline and a refresh schedule.
- Profile audit. Validates an Instagram profile with a 10-point scorecard (0 to 2 each, out of 20), finds the visit-path break point, and rewrites the name field, 3 bios of 150 characters or fewer, link order, highlights plan and 3 pinned posts (explain, prove, sell).
- Reel. Writes any Instagram reel from what the owner actually asks for.
- Reel broll. Writes faceless text-on-screen Instagram reels over B-roll, with one stock-searchable shot per line, cover, caption and CTA, by running the reel job in the B-roll text format.
- Reel long. Writes talking or voiceover Instagram reel scripts of 30 seconds and up (hook, tension, value, proof, CTA, re-hooks, B-roll cues) by running the reel job in the scripted format.
- Reel short. Writes 6 to 12 second Instagram reels with one hook, one payoff and a loop-friendly ending, by running the reel job in the short loop format.
- Slideshow story. Writes photo-slideshow storytelling posts for Instagram where the product arrives late, in three shapes (timeline of 6 slides, glow-up of 4, before/after of 4), with a spine sentence test, one caption line per photo, shot descriptions and a payoff quoted from someone else only when the quote is real.
- Stories. Writes Instagram story sequences of 3 or 5 frames for one purpose (teach, relationship, research, proof or convert), each with frame copy, a visual matched to how the speaker shows up (talking to camera, voiceover or text only), a sticker that teaches the owner something, and a CTA that follows the trust ladder (no paid ask before day 3), plus an optional 7-day story bank.
- Text hooks. Writes and ranks text hooks for Instagram (reel overlays, carousel covers, caption first lines, story first frames) across 12 hook families, sized to each surface, scored on a 5-point rubric with a variety check and a top 3.
- Visual hooks. Designs first-frame visual hooks for Instagram reels across five levers (motion, contrast, object, text placement, pattern interrupt), as on-camera openings when the speaker shows their face and as no-face openings for voiceover or text reels, each with a shot description, the first 1.5 seconds timed, text position, and how it pairs with the text hook.
- Weekly plan. Plans an Instagram week of 3, 5 or 7 feed posts plus daily stories, by pillar and format (short reel, B-roll reel, long reel, carousel, slideshow), with a sales ratio of about 3 in 10, hooks, goals, CTAs and the Nexin job that produces each slot, as a calendar table with optional CSV.
- Account setup. Checks the one-time Pinterest account setup that decides what the account can do and how far Pinterest trusts it: business account, claimed website and stores, rich pins, analytics, country and language, business details, security, and conversion tracking (UTM, tag).
- Analytics. Reads Pinterest analytics and turns them into next moves: what each number means (impressions, saves, pin clicks, outbound clicks, click-through rate, engagement, top pins, audience, conversions), a diagnosis for the common patterns (no impressions, impressions without saves, saves without clicks, clicks without sales, a sudden drop), the pattern the top pins share, and the weekly, monthly and quarterly reviews, with a KPI row to log.
- Board architect. Designs a Pinterest board structure of 10 to 15 boards named as searches people type, ordered on the profile from closest-to-offer to broadest, each with a keyword-led description (max 500 characters), a role (core, supporting, seasonal), the URLs it feeds and starter pin ideas; also sorts existing boards into keep, rename, merge or archive.
- Image prompts. Writes image-generation prompts for Pinterest pins that read as real photos: a concrete scene, named camera and light imperfections, plain materials, no beauty words, no text in the image, and people only as the speaker file allows (no faces by default: hands, backs, objects), composed for a 2:3 frame with clear space for an overlay added later.
- Keyword cluster. Builds Pinterest keyword clusters for a brand or a single URL: pillar terms, supporting terms and long-tail phrases, each tagged with intent (learn, compare, plan, buy), mapped to a board and a destination URL, and labeled verified or unverified depending on whether a search tool was used.
- Niche check. Validates a niche or the brand's current topic for Pinterest before months of pins go into it: demand seen in Pinterest's own search, buyer intent, how it can be monetized with the brand's offers, how long it lasts (evergreen, seasonal or trend), how crowded the top results are and whether they link to products, and the trend direction.
- Pin batch. Writes a batch of 5 to 10 search-first Pinterest pins for one destination URL, built from 3 to 5 distinct concepts: keyword-led title (max 100 characters), natural description (max 500), text overlay of 8 words or fewer, alt text, target board and a 2:3 (1000x1500) design note, with counts computed and a spacing plan.
- Product fit. Validates a product and its page for Pinterest traffic before the pins go out: whether the product suits Pinterest buyers (visual, searched with intent, specific, evergreen, priced for the path), whether the page keeps the pin's promise (matching headline, benefits before features, a real preview, one true specific detail, honest price and urgency, refund and license visible, one next step, fast and working on a phone, an email step), the funnel (direct or through a freebie), a plan of 8 to 15 pins over weeks, and bundle and next-sale ideas.
- Product pins. Writes "what's inside" Pinterest pins for a digital product or freebie page: inventory from the real contents, 3 to 5 concepts (preview, page spread, problem to tool, who it is for, use-in-context), titles max 100 characters, descriptions max 500, overlays of 8 words or fewer, alt text, board, mockup notes and the funnel CTA, with no invented results, reviews or prices.
- Profile audit. Audits a Pinterest profile against a 10-point scorecard and returns fixes: display name with a keyword, username, about text (max 500 characters) ending in the funnel step, profile and cover image, board order, claimed website, business account, link destination and the path from a pin to the freebie.
- Seasonal calendar. Plans a Pinterest pinning calendar that starts seasonal topics about 45 days or more before the moment, sets a cadence the owner can sustain, rotates destination URLs so none repeats the same day, reserves slots for fresh pins on old URLs, schedules design batch days, and exports a date table (CSV or JSON on request).
Copy
- Ai tell scrubber. Finds the patterns that make text read as machine-generated (em dashes, balanced "not X but Y" pairs, stacked triads, throat-clearing openers, summary closers, filler words, uniform rhythm, vague intensifiers) and rewrites each one so the copy gets more specific and more like whoever speaks for the brand, not just safer.
- Angle pool. Builds a pool of message variants for one offer, freebie or topic, sized to the posting cadence and spread across six angles (reassurance, mechanism, count, payoff, constraint, objection), so the account can post for weeks without repeating itself.
- Cta ladder. Writes calls to action matched to how much the reader trusts the account, per channel and per day, from soft asks (save, reply, follow) through freebie asks (comment keyword, link in bio) to direct offer asks.
- Editor. Final craft pass on any finished draft (post, batch, chain, carousel, script, email, page) that tightens the first line, rhythm, specificity, cuts and the ending, keeps the speaker's voice, and returns the edited copy with a short flag list of at most five items for anything it could not fix alone.
- Framework router. Chooses the copy framework that fits the reader's awareness level and the format (PAS, AIDA, BAB, 4Ps, Hook-Story-Offer, PASTOR, Star-Story-Solution or a step series), explains the choice in one line, and writes or rewrites the piece with every stage labeled.
- Headline lab. Writes and scores headlines, email subject lines with preview text, video titles, blog titles, pin titles and sales page headlines, rating each for clarity and curiosity on a 0 to 5 scale, checking length and keyword placement per format, and returning a ranked top 3 with the reason for each.
- Hook lab. Writes a varied batch of opening lines (hooks) for any channel from thirteen hook families, each with when it works, the trap it falls into and a bank of fill-in hook templates (mostly for Instagram reels), then scores every hook on a 7-point rubric and returns a ranked top 3.
- Natural speech. Rewrites copy so it sounds like the speaker saying it once, out loud, to one person: plain words, the speaker's own roughness, specific moments instead of "most people".
- Objection bank. Collects the reasons a reader will not follow, sign up or buy, sorts them into eight objection types, and turns each into three ready assets, a post angle, a short reply for comments and DMs, and an FAQ answer, using only proof from the ledger.
- Storytelling. Turns a lesson, tip or offer into a short story using four arcs (micro-story with scene, tension, turn and lesson; origin; transformation; mistake), sized for posts, chains, carousels, scripts and emails.
- Swipe teardown. Takes apart a piece of content the user admires (a post, thread, carousel, script, email or page they paste) across eight components (hook family, promise, structure, proof type, format, emotional turn, CTA, length), explains why it works, sorts what to take and what to leave, and writes an adaptation in the brand's own voice and angle that keeps the pattern but copies none of the words.