AI Money Cases · 2026-10-21

Issue #10: 2 AI rendering products from the same founder, both payment-verified, one grown on SEO and one on cold email only

2 cases this issue, both with payment-platform data on TrustMRR and both from the same founder, Polish software engineer Piotr Obidowski, who builds alone after work: Visualizee, an AI rendering tool for architects and interior designers, at $8,394 MRR; it made about $150 a month for its first two years and took off after he swapped a node editor for a chat box that asks follow-up questions and moved to monthly subscriptions, with SEO as the main channel; and VizTunr, a try-on button for car tuning shops built on the same rendering engine, at $791 MRR, where the founder reports about 1,000 cold emails brought in 5 paying customers, each email being just a link to a page already built for that shop (customer and email counts are self-reported, unverified).

Cases

SaaS

Visualizee: AI renders for architects and interior designers

MRR$8,394ⓘ

Business model: B2B SaaS subscriptions: 5 free watermarked renders; Hobby $15/month, Pro $35/month, Max $80/month, 25% off yearly; paid plans have no watermark and include a commercial license (pricing page, 2026-10-21). Payments via Stripe (TrustMRR syncs through the Stripe API).

Credibility★★★★4/4Difficulty3/5
See how it grew →
SaaS

VizTunr: an AI try-on button for tuning shops

MRR$791ⓘ

Business model: B2B SaaS subscriptions, one per shop, priced by render volume: Starter $49/month for 100 renders, Growth $99/month for 250, Pro $199/month for 550; every plan includes every feature and unused renders roll over (pricing page, 2026-10-21); there is also a pay-per-render Personal Studio. Payments via Stripe (TrustMRR syncs through the Stripe API).

Credibility★★★★4/4Difficulty3/5

Tool pick

Claude Code

Strong at long tasks and repo-wide refactors; the most frequently mentioned tool in our cases.

Takeaway

The same rendering capability, sold to two kinds of customers, grows in completely different ways: architects search for "sketch to render" themselves, so Visualizee runs on SEO; tuning shops don't search for a tool like this, so VizTunr has to go to them, and builds the product for them before sending the email. Both show that users are often held back not by AI quality but by a product that is too hard to use (the node editor) or by not being able to picture the result (how the part looks once fitted).

Submit a product

Submitting does not guarantee inclusion. We check the original posts.