Skip to content
MakerGain
← Back to cases

Upload a screenshot, choose a tone: how a reply script became a subscription business

Edited by MakerGainPublished: Article updated:

Social Wizard suggests what to say next when a chat stalls: upload a screenshot, choose a tone and get a few possible replies. Founder Kelechi Onyeama first wrote a script for himself, then turned it into a product. Invited to homecoming by a friend, he let several people try it before launching. His repeatable format was uploading a Story screenshot to generate a reply, then paying micro-creators to demonstrate it. He recalls one video bringing tens of thousands of US dollars in revenue. Downloads include three free days, followed by a subscription.

Someone replies “haha.” You type three characters, delete them, try another line and delete that too. Social Wizard takes over that moment: upload a chat screenshot, choose one of ten tones, and receive several possible replies. Its product page calls the output “hints” and “reply suggestions,” without specifying whether it can press Send for you. Users can also upload a photo of a person or event, choose a tone and get an opener, or enter situational details without an image. Every downloader gets three free days; continued use then requires a subscription.[1]

Image: official product demonstration[2], obtained on 2026-11-18, retaining the original presentation for product illustration only.

His own problem came before the product

In the October 28, 2025 interview, founder Kelechi Onyeama said the first version was not a product. It was a NestJS script generating replies for himself.My 2 apps made $1.5M - Starter Story[3] That order deserves a pause: someone got stuck in his own chat window before writing code. The interview introduction describes a 22-year-old from Nigeria who arrived in the US with 100 US dollars. His personal website says he built Social Wizard while homeless.[4]

A friend invited him to homecoming, where he let a few people try the app. He said it was probably the first time his high-school friends cared about something he had made. Further refinement and the App Store launch followed.

The material it handles makes privacy an explicit issue. Uploads may be chat screenshots or photos of people and events for openers; screenshots may include messages from others, with no public figures on how often that happens. The public privacy policy, updated January 13, 2024 and naming Odanta LLC, provides only general retention rules: information is kept as needed for its stated purposes, then deleted or anonymised when there is no ongoing legitimate business need. Backups that cannot be deleted immediately are securely stored and isolated until deletion is possible.WIZARD PRIVACY POLICY[5] The document does not specify screenshot handling, retention days or training use. App Store privacy labels list data types only.

What he replicated was the format

He first made content himself, looking for a repeatable format. The one he found was uploading a girl's Story screenshot and generating something to reply. The action is inherently visual: see a Story, capture it, put it in the app and turn it into a reply.

He then took that format to micro-creators who had not had a truly viral moment, though some videos had reached 100,000 or 200,000 views, and paid them to perform it.

One specific example: he recalled paying a creator about 120, with no currency stated in the original. The video reached 2 million views and brought tens of thousands of US dollars in revenue. The interview supplies neither video analytics nor settlement records. He considered it worthwhile and said he then repeated the format with multiple creators and videos. The public interview does not itemise each promotion fee, settlement or return, so there is no checkable evidence that the other creators and videos were equally worthwhile.

His revenue timeline was a first sale in late January for about 10, currency unspecified; roughly 3.5K MRR two weeks after finding the viral format; over 25K MRR a month later; cumulative receipts of about 250K by month six and about 500K by year-end. The spoken amounts in the interview carry no currency. These are reported MRR and cumulative receipts, not profit. In the same interview, he said two apps together generated 1.5 million US dollars over the previous 12 months. The other was CleanEats; that combined figure does not all belong to Social Wizard.

The video introduction places nearly 1 million US dollars in revenue directly after Social Wizard, while the interview body discusses 1.5 million US dollars across two apps. Those are different scopes. Public materials do not clarify the period or reconciliation for Social Wizard alone. His personal website reports $800K+ revenue for Social Wizard without a period, 700K+ downloads, a 4.8-star rating and 6.6K+ US ratings, and says it now runs passively while still earning revenue. The top of that page lists $254K a month across his apps. It names Social Wizard, the sold CleanEats, Magneta and Caspade—not one product. Those revenue figures also lack platform records or a third-party audit.

Weekly costs the most, but the weekly payer count is unknown

In the interview, he quoted three prices: about 10 weekly, 20 monthly and 80 annually, without naming a currency. The App Store purchase list is more varied: Unlimited Wands appears repeatedly at $9.99; $19.99 covers both Unlimited Access and Unlimited Wands Monthly; both $49.99 and $79.99 are called Unlimited Wands Annual; there is also Threads at $2.99. The page does not identify the current default tier, and some entries omit durations entirely. Public purchase entries therefore establish neither the current default plan nor every entry's billing period.[1]

At the interview's weekly price, four weeks cost about 40, twice the monthly price. Public materials do not say how many people actually pay weekly.

He made two cost statements. One was that marketing accounted for most spending, margins after Apple's commission exceeded 90%, and infrastructure cost about 1–2K monthly, with currency unspecified. The other was that he had not spent one US dollar on marketing in the latest year while the app kept earning. The interview separates early marketing expenditure from that recent no-spend year, but provides no period-by-period spending, cost coverage or complete bills. The over-90% margin is his claim, not a basis for calculating net profit.[3]

Another product approaches the same stuck moment differently. SLIDD's website describes an iOS keyboard that reads the current screen and inserts a reply with one tap, without screenshots. It supports Hinge, Tinder, Bumble, Instagram, WhatsApp and iMessage. After a three-day unlimited trial, prices are $6.99 weekly, $19.99 monthly or $99.99 annually, with weekly preselected. It is currently iOS-only in the US and Canada.AI Assistant for Dating Apps: One Tap, No Screenshots[6] Both address that same blank moment; the distinction is one fewer screenshot round trip. These are SLIDD's own claims. There is no like-for-like comparison or migration data.

What was truly replicated in this story appears to be the video-ready format rather than the model. It made “what should I reply?” into an action that could be demonstrated and repeated, then attached a subscription.

Continue reading

AI companionsAI imagesAI audio
One reply counts as one, a photo as eight: Embraces.AI breaks companionship into messages and sells it

Embraces.AI prices companionship by message: one reply counts as 1 message, and one Picture this image counts as 8. Companions have cross-session memory, send multiple bubbles in a row, and initiate contact; users can edit memories and export chats. Before signing up, there are 40 free messages; the three paid tiers start at $12, and $5 tops up 250 messages. TrustMRR recorded on 2026-11-14 all-time revenue of $48,193 and 426 active subscriptions.

AI publishing
After switching off Claude Opus: AI Flow Chat puts content systems on a whiteboard—and usage bills on the app owner

AI Flow Chat lets creators put videos, PDFs, websites and other references on a whiteboard, group them to define a style, and build a repeatable content system delivered through scheduled tasks or CMS embeds. Embedded-app usage is charged to the owner. After a VC-backed startup failed, founder Alexander Van Le turned to bootstrapping, disabled Claude Opus because it cost ten times as much, and replaced three usage tiers with a single credit system to control model-call costs.

AI audio
Someone will take over the work after you record a podcast—for US$3,000 a month

Two founders take over the entire post-production workflow after a podcast is recorded. Creators upload material and review it each week; the team handles revisions, publication and scheduling. Execution moves to the service, while judgment stays with the creator. One founder quit his job to build a technical product, judged it “not a good business,” then moved into this service and writes the audio pipeline. The other ensures each episode is delivered on time.

AI productivity
Over 180,000 views, 1 February sale: AI Renamer leaves batch-renaming confirmation to people

Ozgur Ozer's AI Renamer gives content-based names to opaque filenames such as DSC_0412.JPG and scan_20250312.pdf. For manual batches, people inspect the suggested names before pressing Rename All; nothing changes on disk beforehand. Ozer had always given away his small apps and first used Polar to solve payment collection because Turkey had no Stripe support. His first post exceeded 180,000 views but produced 1 sale that month; sales began coming after local support and a Product Hunt launch.