MakeViral, a one-man app turning TikTok memes into one-tap videos, jumped 2.6x to about 4,400 monthly visits in June, then slid to 3,200

Traffic moves in meme waves, not a compounding ramp

makeviral.ai is a 3.2K-visit-a-month site. That is the September 2026 estimate, and it comes with a global rank of #5,769,134, a 19-second average visit and a 38% bounce rate2 — the shape of a tool people open, make one video in and close. Every figure in this chapter is a modeled third-party estimate, and at this size the two models disagree badly: for the same September, one puts organic search at 32% of visits (about 1,030) while the other counts 154 organic visits for the whole site in that month218. Read the direction of the curve, not its decimals.

The curve has ten months of data, and it opens with three zeros. October, November and December 2025 show no modeled traffic, which measures the panel's floor rather than an empty site: the domain was registered on 11 November 2024, the founder dates his work on Makeviral.ai from August 2025, and a video-hosting subdomain appears in certificate records on 5 November 202531009. First modeled traffic is January 2026 at 1,465 visits, then 1,665 in February2. Nothing dated explains that first appearance; the public traces all run two to fourteen months ahead of it.

The first real swing is March, when visits rose to 2,292 and social jumped from 9.5% to 49.7%2.7% of them, the only month in the series where social outweighed search2. That month has no dated trigger — no launch post, no ranking change, and the name searches on Hacker News and Indie Hackers return only different tools with similar names721132. It also did not hold: visits fell to 2,197 in April and 1,692 in May2.

June is the largest jump in the series, 1,692 to 4,391 visits, up 160% in one month, and the search estimate roughly tripled from 43 to 1,184 in the same window218. Two independent models rising together suggests something happened, but nothing on file accounts for it. The video-model releases that often get blamed sit in February, when Kling 3.0 and Seedance 2.0 shipped112117, and the template and SEO build-out shows up only in September and October, at 92 and 352 pages updated1. The bump then behaved exactly like a trend wave: the search estimate fell from 1,148 in July to 58 in August and 53 in September while total visits held near 3.2–3.9K182. Two counting quirks make that stretch harder to read — from April to July the provider reports 0% for every channel while still counting visits, and full channel detail only returns in August, so monthly channel splits in that window describe the model, not the business2.

The one spike with a documented cause is October 2026, and it is the site's operating model in miniature. The stop-sign meme came from real police bodycam footage in May 2025; an AI Chihuahua version was posted on 5 October, CapCut tutorials showing it passed 1.6 million views on 6 October, Google searches for “what stop sign ai” started on 7 October, and MakeViral published its template on 18–21 October50. The search estimate moved from 53 visits in September to 1,157 in October18. Ten days between the search wave starting and the page existing is the fastest trend-to-traffic link in the data; the October visit total is not in the series yet, so the spike shows up in search only.

Scale is what keeps this in perspective. In September 2026 crayo.ai modeled 985K visits95, revid.ai 911K19 and autoshorts.ai 177K20 against makeviral's 3.2K9519202. Over the same twelve months revid fell about 45%19 (1.67M to 911K), autoshorts about 23%, and crayo rose from 700K with a spring trough, so the category was flat-to-soft rather than surging192095. None of the three shows a June jump of 160%192095. makeviral's moves were its own, made at roughly one three-hundredth of crayo's size.

No launch: a private domain, then nine quiet months

The first public trace of makeviral.ai is a domain record, not a post. The domain was registered on 11 November 2024 with a private registrant3, and certificates for the bare domain and for www were issued the same day9. Nothing else dated turns up for the next nine months: no capture of the site, no announcement, no listing9100.

The product's first dated statement is the founder's own. İrfan Şener's LinkedIn dates “Founder - Makeviral.ai” to August 2025, with the line “Just started to working on new startup to help people to go viral on Social Media while promoting their products”100. That pitch — promote your product while going viral — matches the URL-to-video tool still on the site, which takes a product link and renders a vertical video from it88. The next build trace is a certificate for videos.makeviral.ai on 5 November 20259, a subdomain named for video, about a quarter after the founder's stated start date100.

The launch left no numbers behind. Hacker News, Product Hunt, Indie Hackers and Reddit searches turn up no launch post for makeviral.ai and no votes or comments attached to one; Hacker News' public search returns seven stories containing the string but zero Show HN projects7. The named products that search does surface belong to other people: MakePostAI, posted to both Hacker News and Indie Hackers in July 2025, is a separate AI-posting tool21132, and Product Hunt's “Make Viral Animations” is AutoAE's23.

Traffic begins in January 2026 at about 1,465 monthly visits2. The two months after — roughly 1,665 in February and 2,292 in March — keep the site at side-project scale, not launch scale2. That first month split 46.8% search and 295% direct2, with 109% referrals2, though at a thousand visits a month the mix is indicative rather than precise. Between the August 2025 start date and the first modeled visitor there are roughly five months with no measured audience at all2100.

There is no verifiable count of first users. The only figure the site publishes is “11,000+ creators use MakeViral.ai”, set beside “245 trend templates” and “1.9 billion views on the trends we cover” on its affiliate page98. Its reviews page says it publishes no testimonials, star ratings or review counts because it has not collected reviews it can verify22, and the launch and review sites that a first-year product would normally sit on return no listing with a rating for it723. The 11,000 figure cannot be dated to the first months or checked against anything outside the site98.

Who was behind it is the one thing the early record settles. Searches for “makeviral” on Hacker News, Product Hunt and Indie Hackers turn up other AI-content tools rather than this one72123, the GitHub repositories of that name belong to other projects92, and İrfan Şener lists makeviralai among his projects100103106. The record before October 2026 therefore reads as a private domain, a founder's LinkedIn entry and a video subdomain; the first third-party listing, in a ChatGPT directory, dates from that month77. Everything measurable about its audience starts in 2026.

Creators sell it; search barely does

Each row below is a slice of about 3,200 visits in September 2026, and at that size the mix is a shape rather than a count2. Direct (34.6%)2 and organic search (32.2%) run level, with social, referrals and email in single digits behind them2. The two providers disagree by an order of magnitude on the search row itself — about 1,030 visits against 154 for the same site — so every share here is indicative218.

Direct, roughly 1,100 visits a month, is the biggest single bucket, and it looks like a product that spreads as clips: the visitor already has the name and arrives with no referrer2. The only brand query in the related-terms list, “makeviral”, carries about 180 searches a month, while the larger related terms are generic TikTok questions — the name appears to travel by share rather than by search2. Who those visitors are is unsettled: one provider reports Turkey 29.7%2 and Vietnam 25.1%2, the other the US 41% and India 29.9%, so no channel can be pinned to a country218.

Search is the channel with the most pages and the least payoff. The whole organic channel is counted at 154 visits a month, with the homepage taking about 45 of them and the stop-sign template about 1118, which makes the 71.4% homepage share a share of a rounding-sized base18. The keywords behind it are four phrasings of “how to make your video go viral” at #1 and “stop sign meme” at #618. Domain Rating 51 on 372 referring domains says the links exist; the 444 URLs in the sitemap, mostly template and tool pages, have not turned into a long tail118.

Social is 9.1%, about 290 visits2, after one month — March 2026 — when it briefly took 49.7% of 2,292 visits2. The deliberate machinery behind that channel is creator-led: an affiliate program paying 20% of every payment for 12 months, links tracked for 60 days, a $50 payout floor, aimed at TikTok and Reels creators, meme and edit pages and AI-tool reviewers98. It is a program rather than a proven channel: no named creator, tutorial or directory that demonstrably sent users surfaced, and the indexed Vietnamese and Turkish videos turn out to be tutorials for other tools83. The site's own banner claims 11,000+ creators, against about 3,200 visits a month modelled for the site298.

Below that sit the cheap entry points. Referrals are 8.2%2 and email 4.7%, and inside them runs an invite loop paying +100 credits for a first paying friend and +50 for each one after, with +50 for the friend2109. AI assistants send 2.0%, about 64 visits, all from chatgpt.com2.com2, through a ChatGPT app created in October 2026 whose directory listing scores 55 against 79 for leaders77 and an MCP server that lets Claude or an agent run templates in-chat72. Paid search (1.3%)2 and paid social (1.1%) add up to about 80 visits2, against 26.9% paid search and paid social combined at a peer faceless-video site, autoshorts.ai, takes 25.6% of its traffic from paid search alone20; no ad campaign for makeviral appears in the material.

The rows together describe a pull that starts outside the site. A meme takes off, the one-tap template page appears within days — the stop-sign template landed about two weeks after Google searches for the phrase began — and a creator with an affiliate link shows the finished clip to people who already know the trend, so they type the name or tap the link instead of searching5098. That is why direct and social outrun search and why 444 pages buy so little of it. The timing of each spike belongs to the growth curve; the channel logic is that the product is sold inside the trend, not in a keyword.

The homepage still sells the 2025 product

Makeviral's front page and the rest of the site describe two different businesses. The homepage is titled “MakeViral - AI-Powered Viral Short Video Generator”, headlines with “Turn Any Product Into Viral Content”, and walks you through pasting your product URL so the AI can write and generate ads for it, promising 10x faster creation, 95% cost reduction and a counter reading 2K+ videos created4. One click inside, the catalogue is 263 trends priced per tap — 123 pet templates, 82 dance, 78 funny, 62 cat, 59 dog — and a faceless video costs 10 credits while a trend template costs 26 to 30458.

The earliest positioning statement on record matches the homepage, not the catalogue. In August 2025 the founder dated the project and described it as a startup “to help people to go viral on Social Media while promoting their products”, the same double promise as the front page100. The domain and its www certificate were issued on 11 November 2024, a video-hosting subdomain appears on 5 November 2025, and modelled traffic starts in January 2026239. No archived copy of the homepage exists, so its copy cannot be dated to a month; the trend catalogue carries no launch date either, which leaves the sitemap's update months as the only dated trace of the build-out.

What the site does document is a feature set assembled around templates rather than around product URLs. The sitemap submits 444 URLs, 92 of them updated in September 2026 and 352 in October 2026, including 31 template pages, 8 tool pages and 6 blog posts1. Pricing lists faceless video, slideshows, text-conversation ads, “AI creators” and photo tools beside the templates5; the reviews page states plainly that there is no free plan and no posting or scheduling22; a REST and MCP server puts the same templates inside ChatGPT and Claude72. Each trend page also doubles as a record of the meme: the stop-sign page runs a dated timeline from the May 2025 body-camera repost (2.Each trend page also doubles as a record of the meme: the stop-sign page traces the joke to real body-camera footage of a Daytona Beach traffic stop in May 202550.

The one change that meets a movement in the data is the trend calendar, and it arrived late. Visits peaked at about 4,391 in June 2026, months before the September–October page build-out, so no positioning change explains that rise2. Search estimates for the same site moved from 43 visits in May 2026 to about 1,184 in June, 1,148 in July, 58 in August, 53 in September and 1,157 in October18. The October rebound lands in the same weeks as the 352 pages updated in October, and the two cannot be separated118; the growth-curve chapter has the visit series.

The build-out did not turn into long-tail search. The homepage carries 71.4% of modelled top-page organic traffic, the top keywords are generic how-to phrases (“how to create viral videos” at #1) plus one meme term (“stop sign meme” at #6), and the whole organic channel is estimated at about 154 visits a month18. The sitemap submits only 31 template URLs against the 263 templates the catalogue advertises, so most of the product has no submitted page at all18. The pivot was a product change first and an SEO play second.

Speed of model adoption is visible in pricing but not measurable as an acquisition moment. The list labels Seedance 2.5 “NEWEST”, Veo 3.1 “PREMIUM”, Wan 3.0 “NEW” and Grok Imagine “CHEAPEST”, with Kling 3.0 Omni alongside them5; Kling 3.0 released on 6 February 2026 and Seedance 2.0 on 8 February120129, but no record dates when MakeViral added either, so per-model lag cannot be computed. What the traffic does show is that the peaks follow memes, not releases: February and March were ordinary months while the June peak had no release behind it2, and the October stop-sign wave — Google searches from 7 October, the AI dog clip on 5 October — got a template within about two weeks50. Four near-identical sites published their stop-sign guides in the same 23–27 October window, so the shared method was speed, not language54576670.

The free tier ends at a still image

There is no free video at makeviral.ai. New accounts start with 0 credits, there is no free plan and no trial, and the smallest first purchase is a one-time pack of 100 credits for $9.99221085.99225. The free layer sits one step away from the product: utility pages such as the fake text message generator run in the browser with no sign-up and export a PNG, and that page points readers to the paid generator that animates the same idea into a 9:16 video96. Free ends at the still image; motion is what you buy.

Everything is priced through one credit balance with several ways in522. One-time packs run from $9.9922.99 for 100 credits to $12922.99 for 1,900; the plans are $922.99 a week for 130 credits (first week $3522.99), Starter at $4922.99 a month for 700 credits (first month $18522.99) and Pro at $19922.99 a month for 3,000, with yearly versions at $49922.90 and $1,99922.9022. Per credit that is roughly 10 cents at the entry pack, 7.7 cents on the weekly plan as the page itself states, and 6.7 cents on Pro522. The largest pack breaks the ladder: $12922.99 for 1,900 credits works out to about 6.8 cents, so someone who pays $130 once gets within a hair of the Pro plan's unit price without committing to anything22.

What a clip costs is set per generation, not per month. A trend template video is 26 to 304 credits depending on the template, a faceless-format video is 10, a template photo is 3522. So the $9522.99 pack buys three of the cheapest trend videos and not one of the most expensive, which at 304 credits is about $30 at pack rates or about $20 at Pro rates522. This is a pass-through model: the pricing page says spent credits cannot be refunded “because we pay the AI providers for every video”, while failed jobs return their credits by themselves5. The margin is the spread between the credit price and the model cost, and the customer carries the cost of choosing an expensive model.

Two details make the subscription the less generous half of the mix. Plan credits are spent before pack credits and expire when the plan ends, while pack credits never expire5. The yearly plans are also a flat 17% off paying monthly — $49922.90 a year against $4922.99 a month, $1,99922.90 against $19922.9922. A one-time buyer keeps everything they have paid for; a subscriber who stops keeps nothing left in the current cycle.

The nearest thing to a trial is a discounted first period. The weekly plan starts at $3522.99 instead of $9522.99 and Starter's first month is $18522.99 instead of $49522.99, both renewing at full price5. Testing the product for the minimum therefore costs $3522.99 for seven days. The two entries are not symmetrical either: a forgotten weekly plan bills about $519 over a year, more than the Starter yearly plan at $499522.90522.

Acquisition is paid out of the same margin. The affiliate programme gives creators 20% of every payment for 12 months — $298.00 a week on the weekly plan, $10 a month on Starter, $9998.98 on the Starter yearly payment, $898.00 on a 450-credit pack — tracked on a 60-day link with a $50 minimum payout98. The invite loop follows the same rule: +100 credits for a first paying friend and +50 for each one after, +50 for the friend, released only once the invited account has bought something109. Nothing on the site is free of a payment somewhere, and the referral budget is a fifth of any payment, pack or plan.

What those payments add up to is not public. The affiliate page's “11,000+ creators use MakeViral.ai” is a usage line, not a count of buyers, and the review page says the company publishes no testimonials or star ratings because it has not collected verifiable ones9822. The founder's profile lists the company as a current role with no revenue, customer or funding details100. The business model is therefore legible in its price list and invisible in its results.

One indie app developer, no company on the record

makeviral.ai has a named operator and nothing that looks like an organisation around him. İrfan Şener's LinkedIn lists "Founder - Makeviral.ai" from August 2025 to the present, remote, from Kazakhstan100, and his personal site lists makeviral.ai in the same project folder as his other apps103. The domain was registered on 11 November 2024 with a private registrant through NameCheap3 — nine months before that start date, so the paper trail begins with a domain, not with a company.

His background is mobile apps, not AI video. Apptechn, which he founded in December 2019, is described on the same profile as a mobile-app development company with "over 10 apps" on the App Store and Google Play and more than 4 million downloads100. His own site counts larger — "30+ Apps · 10M+ Downloads" over eight years, with a folder of "30-stealth-mobile-apps" beside makeviral.ai and a RevenueCat-dashboard generator called TrulyBlessed that he says went viral103 — and a directory profile repeats the 30 apps and 10 million downloads over seven years, describing him as "an app developer from Bangkok"106. The gap between 10+ and 30+ looks like a definition — the studio's live store listings versus everything he has shipped — but only the smaller number comes with a named period, and the location is just as loose: Kazakhstan on LinkedIn100, Bangkok in the directory106.

No second person is named anywhere public. Help-centre articles carry the byline "By MakeViral Team" with no individuals attached135174, and the affiliate page promises applicants "a real person on our team" who reads every application by hand98. That is consistent with a sole operator rather than a staffed company, and it matches the site's size: about 3,200 monthly visits in September 20262 and 372 referring domains18. Nothing published states how many people work on the product, so the team size stays an inference from scale and bylines.

The company side is equally thin. Beyond the private domain registrant3, the founder profile shows three concurrent founder roles — Makeviral.ai, Apptechn and an unnamed stealth-mode startup — rather than a board or any mention of investors100. The "11,000+ creators use MakeViral.ai" line on the affiliate page98 is sales copy; creators is a cumulative count and visits are a monthly one, so the two cannot be set side by side, but neither is corroborated by anything outside the company's own pages.

What does come through is that he frames the business as a download-count project rather than a SaaS company. The directory profile says he is "aiming for 100m+ downloads" with makeviral.ai106, and his site's one-line philosophy is "8 years building apps, 10M+ downloads, still hungry"103. For a reader weighing the rest of this report, that matters: one developer with a portfolio habit decides what gets built and when, with no launch committee and no outside timetable to answer to.

Peers are 300 times bigger, and they are shrinking

In September 2026, on the same modeled monthly-visit measure, crayo.ai drew about 985,000 visits, revid.ai about 911,000 and autoshorts.ai about 177,000, against makeviral's 3,2002192095. That is roughly 300 times smaller than the two leaders and 55 times smaller than the third — and these are the very tools makeviral writes its comparison pages against3541. Whatever else is true about its growth, it happened from a base two to three orders of magnitude below the category's incumbents.

The incumbents were not on the way up. revid fell from 1.67M visits in October 2025 to 911K a year later, autoshorts from 230K to 177K, while crayo dipped to 452K in March 2026 before recovering to 985K in September192095. A category whose top is flat-to-declining means a 3,000-visit site's month-to-month swings cannot be explained by a rising tide; where makeviral's own curve came from is another chapter's question.

The leaders are organic-and-direct businesses: in September 2026 revid took 47.9% of visits from organic search and 27.7% direct, crayo 33.4% organic and 46.7% direct1995. makeviral's reported split of 34.6% direct and 32.2% organic reads like the same shape, but at its size one provider counts its whole organic channel at 154 visits a month and another at roughly 1,030, so a matching percentage is not a matching business218.

The products are aimed at different jobs, and makeviral's own comparison pages concede the point: “Crayo's Auto Clip turns a file or a YouTube link into shorts. MakeViral has no clipping feature at all”41. Where it argues instead is price per finished video — $0.67–0.71 against AutoShorts' $1.15–1.46 — and format breadth in place of one automated pipeline35. Its pitch to creators is a template per trending meme, each with a page documenting the trend, backed by a ChatGPT app that sits at #7 for “trend” in that directory7798.

Its clearest win is latency. When the “Stop Sign” meme spread in October 2026, makeviral already had a template page for it, among its top pages by search traffic, and its October organic estimate jumps from 53 to 1,15718. It was not alone: viberot published a “What Stop Sign AI” walkthrough the same month, tryh3 shipped a Topo Gigio template and getstarrd a K-pop jumbotron guide7013584. Speed differentiates, but the race has a crowd in it.

The exposure is that the niche has not shown it produces traffic at all. jumponthetrend.app, built on the same premise, was modeled at zero visits in each of July, August and September 202693. And makeviral keeps no safety net: no free plan and no trial, new accounts start with 0 credits, and the product neither posts nor schedules22, so a cold arrival from a meme search has to pay before seeing output. The affiliate page claims 11,000+ creators and 245 templates98 — a user count that cannot be checked against the visit figures above.