Downside: 'uncensored model' APIs and AI companions. Revenue can look great, but payment channels and platform policy can zero it out at any time.
- Status
- Closed (not recommended: the risk lies in payment rails, app store policy and model-vendor terms, not something the product can solve)
- Why this is written up
- "Uncensored" APIs and companion apps are growing fast; companions even hold steady growth, which looks like demand big labs have ceded. The reason not to follow isn't the demand: payment platforms like Stripe can cut off adult and jailbreak content at any time, Apple and Google are tightening review of companion apps, and distilling frontier models violates model-vendor terms. Revenue can go from one day to zero at any time, with no appeal. The Shannon author valued his own business at only two months of revenue, which shows how long he himself thinks this business can last.
- Confidence
- Medium. Embraces and mion6vpjy8 have healthy revenue; the reason for this entry is risk, not a lack of money. Conditions that would overturn this (when it becomes doable for a small team): payment platforms or app stores issue explicit policies allowing this kind of content, and a small team goes 6 consecutive months on indie without its rails being cut; or companion products find a non-adult vertical (speaking practice, companionship for the elderly) with comparable revenue.
Evidence
Grouped by store. Each group starts with the charts that can be graphed (real curves, data dates below the chart), then the scanner's original notes and source links.
Indie product revenue (TrustMRR)3 evidence · 2 products with curves
- indie (11-02 check): Audn (audn-corporation, "one OpenAI-compatible API connecting four uncensored models", prepaid credits) month by month August $4,388 → September $13,104 → October $36,238, but by week it's $9,861 in week 10-12 → $5,559 in week 10-19 → $1,570 in week 10-26; 30 days $28,984, MRR $0. The report recorded $35k and a 90% drop over four weeks.
- indie (11-02 check): Shannon Lab (shannon-ai.com, self-described as "distilling Claude Opus + GPT Pro into uncensored Kimi / Mistral") $9,596 over 30 days (+70%), 659 subscribers, October $9,727; but the author's listing ask is only $21,000, roughly two months of revenue.
- indie (11-02 core): Embraces.AI (embraces.ai, Singapore, AI companion $12 / $25 / $50) 30-day $16,797 (+35.1%), 370 subscriptions, June $636 → October $17,454, climbing month over month, steady at $3.7–4.0k a week. The revenue itself is healthy.
For-sale listings1 evidence items
- For sale (Acquire, 2026-11-02): mion6vpjy8 AI companion / chat app, one person, 22 months in, $54.5k annual revenue, $41.6k profit, 100% subscription, listed at $120,000 (2.9x profit).
Already doing it
Sellers of Audn, Shannon Lab, Embraces.AI and mion6vpjy8; the same company behind Audn also sells AI pentesting at audn.ai ($999/month, $30 in 47,875 days), which suggests it is already moving toward compliance (inferred).
The hardest step
Payment rails and platform policy, not the product, are the problem.
Who it's for
Not recommended. Doing this for real requires a high-risk merchant route (crypto payments, specialist high-risk acquirers) and legal work, and that is not a small team's strength (inferred).
Should every product builder know this?
No. It is only directly relevant to people doing model resale or companion products, but "the payment rail is a single point of failure" applies to everyone (see [081]).
Change log
Original: Scanner opportunity library #085 (updated daily, 06:00); this page was generated on the 2026-11-08 edition.