How we verify revenue claims

The internet is full of “$100k a month with AI.” What we do differently is publish the evidence, not the conclusion. Cases that fail the bar below either lose their figures or never get listed.

How a case gets published

  1. 1

    Find

    Collect candidates from Indie Hackers, Product Hunt, and similar.

  2. 2

    Check it is alive

    Open the service URL and record the date.

  3. 3

    Identify the founder

    Anonymous lowers the grade.

  4. 4

    Find the money evidence

    A payment screen, public dashboard, or interview.

  5. 5

    Assign a grade

    A–D, based on how many independent records exist.

  6. 6

    Test Korean feasibility

    Payment, tax, and platform limits.

  7. 7

    A person reviews and publishes

    The review is logged.

✓ This is evidence

  • Payment processor screens
  • Public MRR dashboards
  • Figures stated in press or podcasts
  • App store metrics

✕ This is not

  • A revenue screenshot on the founder's own feed, alone
  • Claims on a page selling a course
  • Round-ups with no source

What makes an A — independent records

Records that do not come from the founder. At least one is required before we quote a figure.

  • Figures a third party confirmed via a payment integration — e.g. TrustMRR's Stripe link
  • Metrics the sales platform itself displays — App Store, Google Play, marketplaces
  • Third-party reporting with independent confirmation
  • Investor materials and public traffic data

Supporting only — not enough for an A

  • A revenue dashboard the founder hosts themselves (open startup / /open pages)
  • Payment-screen screenshots posted by the founder
  • The official site — does it resolve, and when we checked
  • Founder interviews — press, podcasts, Indie Hackers
  • Repository activity — GitHub

Interviews sit here for a reason. An editor may select and polish the piece, but the source of the number is still the founder. Someone else retyping it does not verify it. (Moved down from the A tier on 2026-08-14.)

“A revenue dashboard the founder hosts” moved down on the same day. Our A tier means evidence independent of the founder, and an open-startup page is published by the founder on their own server. The list and the definition contradicted each other. It is far better than a one-off screenshot — it stays public, and being wrong stays visible. It is still not someone else confirming it.

What does not count

  • A revenue screenshot on the founder's own social account, alone
  • Claims made on a page selling a course or ebook
  • Round-up articles with no source

An unverifiable testimonial can set tone. It cannot support a number.

The grades

AVerified

Live service + identified founder + at least one record independent of the founder. Figures are quoted.

BMostly confirmed

The business exists, but the only revenue evidence is the founder's own statement. No figures shown.

CSelf-reported

Revenue is claimed and even the business is hard to confirm. Not presented as a success case.

DUnverified

Source unclear or promotional. Not listed.

The same revenue is not the same business

“₩5M a month” is a clear target and an empty instruction. The ways to reach it are different jobs. Below are three A-grade cases we verified, placed by revenue per customer against customer count.

These four are not solo side businesses. Checking each official site (2026-08-14; FrameLiq on 08-15): two are companies with staff and investors, and two have unverified operators. They are used here only to show the distribution of price points — not as things to copy. Both unverified ones are listed for sale ($12,000 and $17,000), which says something about what ends up on a list like this.

₩5M/mo (≈ $3,600)FrameLiq$19 × 12UGC Copilot$35 × 9RankAI$365 × 155AEO Engine$2,084 × 38$10$100$1,000$10,0001101001,000Monthly revenue per customerPaying customers
Values plotted above
CaseComputed (MRR÷subs)Published priceCustomers
FrameLiq$19$19 / $4912
UGC Copilot$35$29 / $79 / $1499
RankAI$365not published155
AEO Engine$2,084$1,597 / $2,99738
Source: TrustMRR (payment-processor-linked figures; UGC Copilot, RankAI and AEO Engine checked 2026-08, FrameLiq checked 2026-08-15). Revenue per customer is our own MRR ÷ paying subscriptions — editorial interpretation, not fact. Three of the four check out against published pricing: FrameLiq's entry plan is $19 against a computed $18.5; UGC Copilot's is $29 against a computed $35; AEO Engine's plans run $1,597–$2,997 against a computed $2,084. RankAI does not publish pricing, so it cannot be cross-checked.

The dashed line joins every combination that produces ₩5M a month. At the left end you need 103 subscribers at $35. At the right end you need two contracts at $2,084. The first is a traffic-and-automation job; the second is a sales-and-trust job. Different skills, different failure points.

This is why every case records price and customer count. A case with only a total cannot be followed.

Distinctions we keep

  • Revenue is not profit. We label MRR, ARR, revenue, profit, and GMV separately.
  • We record what the AI did and what the person did, separately.
  • We record the expertise the founder already had. A ten-year veteran speeding up is not the same as a beginner starting out.
  • “Possible” and “easy” are different claims.

If we got it wrong

Tell us and we will correct or remove it. Contact