Those posting revenue don’t make the top 82; 607 silent websites are collecting money
Max Feng traced referral records from 9 checkout providers and filtered out 607 websites that grow through their products themselves. They overlap with TrustMRR by only 7 companies, and none of the top 82 appear there. Payment-page visits are not the same as revenue, but this list beneath the surface is worth a look.
The several hundred companies posting MRR on TrustMRR overlap with a checkout-provider referral leaderboard by only 7 companies. MRR is monthly recurring revenue; the screenshots look like report cards.
Max Feng and team traced the referral records of 9 checkout providers, including Stripe, Gumroad, and Lemon Squeezy, and filtered out 607 websites that grow through their products themselves. The highest-ranked TrustMRR member on this leaderboard is No. 83. Among the top 82, not a single one is on TrustMRR.
This leaderboard is called TopCheckouts. Its methodology is Similarweb’s estimated number of visits in September from websites to checkout payment pages. If one person clicks twice, it counts twice; opening it is not the same as paying, and it cannot be directly converted into revenue.
How was it filtered? The 9 checkout providers first yielded 15,790 referral websites; they kept 1,344 that had at least 3,000 payment-page visits in September. Then they removed those mainly relying on paid acquisition, leaving 1,280 independent products. After verifying official websites one by one, they kept 608, and 607 made the list.
What lies beneath the surface?
WaveSpeed, an API that calls thousands of AI image, video, and audio models, had its payment page opened 189,000 times per month. Mobbin, a screenshot reference library for app and web design, 146,000 times. Viblo, AI short-video editing, 107,000 times. Turbo AI, which turns lecture recordings, PDFs, and videos into quizzes and flashcards, 99,000 times.
Of the 607 sites, 324 have more than 10,000 payment-page opens per month, and 40 have more than 50,000. The leaderboard splits new-customer sources into SEO 171, GEO 6, social media 79, viral 126, and other 225. Here GEO refers to new customers brought by AI assistants; there are only 6, so AI-assistant traffic is still very small in this sample.
The list is not all AI products. There is Inmatecanteen, which lets people top up commissary accounts for family members in prison; a French vehicle registration certificate agent; online card-pack opening; and a VRChat avatar search directory. They do not post MRR, they do not run in those circles, and their payment pages keep getting clicks.
There are two questions under the tweet worth keeping. One user asked: payment-page opens are still one layer away from revenue, and renewals from existing users and crawlers will both get mixed in—how do you cleanly filter them out? Another user asked: WaveSpeed has no price advantage, so why do so many people use it? The leaderboard does not answer these in the tweet. When looking at the list, you have to know its boundaries: visit counts can show that someone is heading to the payment page, but they cannot tell you how much was ultimately paid.
TrustMRR is the report card above the surface. This leaderboard is the footprints in front of the checkout. Posting revenue is marketing; not posting revenue is business.
Original tweet: Max Feng
发布时间: 2026-10-25 01:30