Three AI wrappers at $17–$119 MRR: dating photos, Reddit leads, and sponsored tokens (2026)

[data] Three one-person AI wrappers cleared this week's filter: each has a founding month in the past 12 months, public traction, and a narrow workflow a small team can explain. TrustMRR supplied the September 7, 2026 snapshots for revenue, team size, founding month, pricing, and listed marketing channels. DatePhotos.AI also publishes its workflow and pricing on its own site. 1 2 3 4

[editor's view] The revenue numbers are early signals. Compare the work each wrapper removes, its distribution evidence, and the risk that remains after the first demo.

TL;DR

[data] DatePhotos.AI turns selfies into dating-profile photos across multiple scenes. TrustMRR lists $119 MRR, six active subscriptions, and $16,198 in all-time revenue, with Stripe verification and a September 7, 2026 update. The listing says the product was founded in May 2026 by a one-person team. 1 [data] MentionCatch monitors Reddit for high-intent buying conversations and helps founders write relevant replies. TrustMRR lists $19 MRR, one active subscription, and $23 in all-time revenue, with LemonSqueezy verification and an August 2026 founding month. 3 [data] sponsored/tokens offers free AI-token access for coding agents, paid for by sponsors. TrustMRR lists $17 MRR, one active subscription, and $70 in all-time revenue, with Stripe verification and a September 2026 founding month. 4

[editor's view] DatePhotos.AI has the clearest consumer package, MentionCatch has the clearest channel-shaped workflow, and sponsored/tokens has the most open business model. All three are cheap to prototype. The harder questions are repeated use, reliable output, distribution, and trust. A first clone test should manually deliver the promised result to five target users before the builder pays for a full automation layer.

Three takeaways

[data] The traction range is $17–$119 MRR. Each snapshot is small, and each product has one active subscription except DatePhotos.AI, which has six. 1 3 4
[editor's view] A listed channel is a starting surface, not an acquisition result. DatePhotos.AI lists SEO and Google Ads; MentionCatch lists Reddit, X, cold DM, and Product Hunt; sponsored/tokens lists no marketing channel. The public pages do not attribute paid customers or revenue to one channel. 1 3 4
[editor's view] The first build should test the job, not the interface. A manual photo package, a hand-curated Reddit lead list, or a sponsor-funded inference concierge can expose repeat demand before infrastructure becomes the project.

Replication scoreboard

[editor's view] The scores run from one to five stars. More stars mean more work or exposure. Technical lift measures the difficulty of producing a reliable workflow. Information edge measures how much private process knowledge or useful feedback can accumulate. Capital needed covers infrastructure, model usage, and distribution spend. Legal risk covers exposure from identity, privacy, platform rules, sponsorship, and user trust.

Product	Narrow workflow	Traction snapshot	Listed acquisition evidence	Technical lift	Information edge	Capital needed	Legal risk	Clone read
DatePhotos.AI	Turn selfies into dating-profile scenes	[data] $119 MRR; 6 active subscriptions; Sep. 7, 2026 1	[data] SEO and Google Ads listed; attribution undisclosed 1	★★★☆☆	★★★☆☆	★★☆☆☆	★★★★☆	[editor's view] Easy to demo; identity and trust carry the risk
MentionCatch	Find high-intent Reddit mentions and help answer them	[data] $19 MRR; 1 active subscription; Sep. 7, 2026 3	[data] Reddit, X, cold DM, and Product Hunt listed; attribution undisclosed 3	★★★☆☆	★★★★☆	★★☆☆☆	★★★☆☆	[editor's view] The lead filter is copyable; relevance and trust are the wall
sponsored/tokens	Give coding agents sponsor-funded token access	[data] $17 MRR; 1 active subscription; Sep. 7, 2026 4	[data] No public primary channel disclosed 4	★★★★☆	★★★★☆	★★★★☆	★★★★☆	[editor's view] The interface is small; sponsor supply, provider terms, and trust are large

[editor's view] DatePhotos.AI gets two stars for capital because one-time image generation can create a direct cost per order, while the product page promises 80–180 outputs and free regeneration. MentionCatch needs less infrastructure, but the useful score depends on finding the right conversations before competitors do. sponsored/tokens needs more capital because subsidized inference creates a bill before sponsor demand is proven.

DatePhotos.AI: the dating-profile photo package

[editor's view] Positioning: DatePhotos.AI sells a fast, one-time way to turn ordinary selfies into a varied dating profile.

[data] TrustMRR lists $119 MRR, six active subscriptions, and $16,198 in all-time revenue. The page says Stripe verified the revenue and shows a September 7, 2026 update. TrustMRR lists May 2026 as the founding month and identifies a one-person team. 1

[data] The official product page says users upload 10–20 selfies, choose Tinder, Bumble, or Hinge, and receive 80–180 AI-generated photos in 20–30 minutes. It describes 40-plus scene variations, a 0–100 realness score, and free regeneration. These are product claims, not independent outcome measurements. 2

[data] DatePhotos.AI lists one-time plans from $29 to $79 and says no subscription is required. TrustMRR lists SEO and Google Ads as marketing channels. The listing gives no channel-attributed customer or revenue data. 1 2

[data] Job-to-be-done: a dating-app user wants enough believable, varied photos to refresh a profile without arranging a photographer or waiting for a conventional shoot. The official page directly frames the workflow around those constraints. 2

[editor's view] The first version needs upload handling, identity-preserving image generation, scene recipes, selection, and payment. The hard part is the rejection loop. A user needs a fast way to discard images that look too polished, inconsistent, or unlike the person in the selfies. A realness score can guide that choice, but the public page does not establish how well the score predicts a user's judgment.

[editor's view] The legal and trust surface is wider than the image API. A clone would handle face images, likeness, privacy, and the possibility that a dating profile presents an artificial life. Dating platforms also set their own rules. A builder should treat those as product constraints to investigate before launch, rather than as a footnote after the model works.

[editor's view] Replication score: Technical lift ★★★☆☆ because the generation pipeline is available but consistent identity and usable selection require iteration. Information edge ★★★☆☆ because scene recipes and rejection data can compound, while the basic promise is easy to describe. Capital needed ★★☆☆☆ because a one-time order can be priced above generation cost, although regeneration can erase margin. Legal risk ★★★★☆ because face data, likeness, privacy, and platform expectations all sit inside the core job.

If you wanted to copy this

[editor's view] First step: sell five manually reviewed photo packages to one dating-app audience, then record which generated scenes buyers keep and which they reject.
[editor's view] First likely failure mode: the output looks impressive in a gallery but fails the buyer's standard for "this still looks like me," causing refunds or no repeat purchase.

MentionCatch: intent detection for Reddit leads

[editor's view] Positioning: MentionCatch turns Reddit reading into a lead queue for founders who want to join buying conversations while the problem is fresh.

[data] TrustMRR lists $19 MRR, one active subscription, and $23 in all-time revenue. LemonSqueezy verified the revenue, and the page shows a September 7, 2026 update. TrustMRR lists August 2026 as the founding month and identifies a one-person team. 3

[data] The product page describes a workflow that monitors Reddit, identifies people actively looking for solutions, scores prospects, and helps founders craft relevant replies. TrustMRR lists pricing at $19 per month, $190 per year, or $199 lifetime. 3

[data] TrustMRR lists Reddit, X, cold DM, and Product Hunt as marketing channels. The page also claims more than 680 qualified leads for four users in less than five days. The claim is presented by the marketplace listing; public material does not establish the denominator's definition, conversion rate, or retention. 3

[data] Job-to-be-done: a founder wants a short list of Reddit conversations where a relevant, useful reply may lead to a customer, without manually searching hundreds of posts. The listing describes that problem and the proposed workflow. 3

[editor's view] The initial build is a feed collector, classifier, scoring rule, and reply assistant. The difficult part is the information edge. A generic keyword monitor produces noise. A useful product must learn which language signals urgency, budget, category fit, and permission to join the conversation. The same user also needs advice that helps them answer without turning a community post into an ad.

[editor's view] The acquisition evidence is unusually close to the product's job: Reddit is a listed surface, alongside X, cold DM, and Product Hunt. The listing still does not show which surface produces paying users. The first test should therefore measure response quality and paid conversions from one subreddit cluster, rather than treating the channel list as a growth plan.

[editor's view] Replication score: Technical lift ★★★☆☆ because collection and ranking are familiar engineering work, while freshness and useful scoring add operational effort. Information edge ★★★★☆ because labeled replies and outcomes can improve the ranking model. Capital needed ★★☆☆☆ for a narrow monitored set; broader coverage raises collection and processing costs. Legal risk ★★★☆☆ because scraping, platform rules, privacy, and unsolicited outreach can all affect the product's trust.

If you wanted to copy this

[editor's view] First step: choose one founder niche, hand-curate 50 recent Reddit posts, and sell five users a daily list with a suggested reply.
[editor's view] First likely failure mode: the tool surfaces posts that contain the right keywords but have no buying intent, so users stop checking the queue.

[editor's view] Positioning: sponsored/tokens tries to make coding-agent inference free to the user by having sponsors pay for the token usage.

[data] TrustMRR lists $17 MRR, one active subscription, and $70 in all-time revenue. Stripe verified the revenue, and the page shows a September 7, 2026 update. TrustMRR lists September 2026 as the founding month and identifies a one-person team. Pricing is shown as free, sponsored inference. 4

[data] The listing's one-line description says the service provides free AI tokens for a user's favorite coding agent, paid for by sponsors. The page gives no model-provider list, sponsor roster, primary acquisition channel, or public usage breakdown. 4

[data] Job-to-be-done: a coding-agent user wants to run inference without paying the provider directly, while a sponsor wants to place a message or fund access in front of developers. The first half is explicit in the listing. The sponsor-side job is an editorial interpretation of the two-sided description, not a disclosed customer result. 4

[editor's view] The narrow prototype is a gateway with usage accounting, sponsor controls, disclosure rules, and a way to route eligible requests. The interface may be small. The business has to answer harder questions: who pays before usage grows, how sponsors are matched to users, what data crosses the gateway, and what happens when a provider changes terms or limits.

[editor's view] No public primary acquisition channel is disclosed. That gap matters more here than in a normal freemium tool because the product needs two sides at once. Developer demand without sponsor demand creates a subsidy bill. Sponsor interest without useful developer usage creates an advertising product with no audience.

[editor's view] Replication score: Technical lift ★★★★☆ because reliable routing, metering, and agent compatibility are operational work. Information edge ★★★★☆ because sponsor pricing, usage patterns, and agent compatibility can compound if the builder gets access to both sides. Capital needed ★★★★☆ because subsidized inference can create cost before sponsor revenue arrives. Legal risk ★★★★☆ because sponsor disclosure, user-data handling, provider terms, and code generated through subsidized inference all require careful review.

If you wanted to copy this

[editor's view] First step: run a manual sponsor-funded pilot with one coding-agent community, one clearly disclosed sponsor, and a hard usage cap.
[editor's view] First likely failure mode: users consume the subsidized tokens faster than sponsors value the audience, leaving the operator with an uncapped model bill.

What to do tomorrow morning

[editor's view] Pick one of these three jobs and recruit five people who already perform it: a dating-profile refresh, a Reddit lead search, or a coding-agent session. Deliver the result by hand, charge for it, and ask each buyer to pay for the same result a second time before building the full wrapper.

Sources

Source	What it supports
DatePhotos.AI on TrustMRR	[data] Traction, active subscriptions, founding month, team size, verification date, pricing, and listed channels
DatePhotos.AI official site	[data] Workflow, product claims, delivery time, scoring, and one-time pricing
MentionCatch on TrustMRR	[data] Traction, active subscription, founding month, team size, verification date, pricing, workflow, and listed channels
sponsored/tokens on TrustMRR	[data] Traction, active subscription, founding month, team size, verification date, pricing, workflow description, and disclosed channel gap

Fuentes de referencia

1
DatePhotos.AI on TrustMRR
trustmrr.com
2
DatePhotos.AI official site
datephotos.ai
3
MentionCatch on TrustMRR
trustmrr.com
4
sponsored/tokens on TrustMRR
trustmrr.com

TL;DR

Three takeaways

Replication scoreboard

DatePhotos.AI: the dating-profile photo package

If you wanted to copy this

MentionCatch: intent detection for Reddit leads

If you wanted to copy this

sponsored/tokens: a sponsor-funded access layer

If you wanted to copy this

What to do tomorrow morning

Sources

Fuentes de referencia

Contenido relacionado