x-octo home Business judgment on AI products
中文

Real-revenue cases

Reclaim

Nice-to-have

An iOS app that helps users quit porn: it blocks adult sites and apps, offers a panic button for moments of urge, and includes an AI companion that knows the user's streak, triggers, and past relapses.

Stripe-verified
Monthly revenue
$272 / mo
Primary source
View original

01

Real-demand verdict

Real-demand verdict

Nice-to-have

It replaces white-knuckling it or generic blockers, handing in-the-moment intervention and relapse history to an AI companion. The quitting need is real, but it's a non-essential category prone to abandonment, so payment retention is uncertain.

How it makes money

Users pay for the iOS app; revenue is Stripe-verified at $272/month. Pricing and billing structure are not stated in the source.

What old behavior it replaces

Relying on willpower alone, or using generic site/app blockers without an AI companion.

02

Where the first customers came from

Not mentioned in the source

03

Tactics you can copy

  1. 01Tie the AI companion to the user's own streak, triggers, and relapse history so intervention happens at the moment of urge, not in later reflection.
  2. 02Add a panic button in a quitting product, turning the highest-need moment into a single clear action.
  3. 03Publish Stripe-verified revenue as a credibility signal for an early-stage product.
Moving it to an AI business

Wiring an AI companion into the user's behavioral data (streak, triggers, relapse history) so it intervenes at the critical moment transfers to any habit-change or health AI app.

04

Evidence and limits

Self-reported numbers are unaudited — treat them as leads, not facts
Stripe-verified

Revenue basis · Stripe-verified (TrustMRR)

What evidence is missing

No pricing, subscription-vs-one-time, or paying-user count, so the unit economics behind $272/month can't be judged.
No retention or churn data, leaving the key question for a quitting product — sustained payment — unverified.
First-customer source and acquisition channels are entirely missing; transferability can only be inferred from product features.