An AI roleplay sandbox game: players build a hex-map world tile by tile, and AI generates settlements, NPCs, art and storylines; NPCs have memory and turn-based combat is available.
A couple (engineer Jacob Garner and his wife, who handles community and marketing) built an AI solo RPG sandbox in their evenings and weekends and launched it in late six months ago. They first posted on their own in Reddit's AI roleplay communities and quickly hit a ceiling (rules banning promotion, downvotes for having no free tier), so they switched to running Reddit ads in D&D and gaming communities, using polished visuals to draw players who normally avoid AI. About 3 months after launch MRR passed $1K and the ads could pay for themselves; afterwards (after three months ago) they added a free tier and last month increased ad spend, and checkout page traffic grew about 7x.One-line growth story, compiled by us from public sources; speculation is marked separately below.
The main engine is paid advertising: it runs targeted Reddit ads in D&D and gaming communities, using polished visuals to attract players who do not touch AI; the second-largest channel is paid social, and landing pages carry Reddit ad parameters.
The lubricant is retention and the free tier: existing players come back to keep building their own worlds, with a bounce rate of only 26% and an average stay of 8 minutes; the free tier added later lowered the barrier after clicking in from an ad.
By estimate, about half of visits come from seeing Reddit ads; the visit count is an estimate, and about 33 clicks to the checkout page per 100 visits, a figure that comes from estimating total visits as checkout page visits x 3.
Model: Lenny's racecar growth framework. Engine = paid advertising (running targeted Reddit ads in D&D and general gaming communities, with subscription revenue buying more ads); lubricant = retention and the free tier (worlds players build are stored on the site and they come back to keep playing; the free tier added later lowered the barrier after clicking in from an ad). The cold start was the founder's own posts in the Reddit AI role-playing communities.
Real data: +151.4% month over month, the largest channel direct at 41.36%, top three channel ranking (direct, paid social, organic search), organic search 100% of search, only 2 search terms (rpbuddy, ai rpg sanbox), country (US 100%), bounce rate 26.41%, 4.63 pages per visit, average time on site 8 minutes 23 seconds and global rank #885821 come from the Similarweb public page, last month; the public page does not give total visits or the share of the second and third channels, and referral and social sources are empty. 38,682 visits (about 38,700) to the Stripe checkout page come from quietmoney checkout page data, versus 5,333 the month before that and 0 the month before that.
Supporting evidence: A couple building it in their spare time with no funding, when it launched in 2026 it only had a 7-day trial plus paid plans; the founder's long LinkedIn post says that about 3 months after launch MRR passed $1,000, ads in D&D communities pay for themselves and the $20 tier is the most popular; Wayback captured a landing page carrying rdt_cid (the Reddit ad parameter); a free tier called Traveler was added later; Discord has about 208 people; Product Hunt launch got 0 votes.
Inference: Total visits estimated from the Similarweb global rank #885821 are about 34,600 (fitted on 437 sites with real visit data, half of the sites within 17% error), even fewer than the 38,700 checkout page visits quietmoney recorded, which does not add up, so total visits instead use checkout page visits x 3 = 116,100, and both channel circles carry ≈; this multiplier is a guess, the error could be a factor of two, and "about 33 clicks to the checkout page per 100 visits" also comes from this multiplier. The channel shares and ranking are real, and how each path is split is an inference: direct 48,000 = returning existing players 41,000 + opening directly after seeing an ad 5,000 + seeing posts and the promo video 2,000; other channels 68,100 = clicking Reddit ads and searching the brand name after seeing an ad 54,000 + existing players searching rpbuddy 10,000 + posts and the promo video 4,100. By ranking paid social is above organic search, so the ads side is given more. The 1% click-through after seeing an ad and 5% click-through after seeing a post are guesses.
Checkout: checkout.stripe.com
“Traveler: Free, no card required. 50 image credits/month and 50 AI actions/day”Quote from the site
New-user source: Other (Direct / referral / email / paid). Paid social 44.0%, direct 41.4%, search 8.8%; new users mainly come from social ads, and the only search terms are rpbuddy and ai rpg sanbox (brand keywords 45.8%). No breakdown of referral or social sources; returning users come back directly to keep playing their own worlds.
Top channel measured by Similarweb, others estimated by rank
| Attributable new users, last month | Share |
|---|---|
| SEO | 4% |
| Other | 96% |
Direct + brand search 46%. Direct and brand search count as viral only when the site shows sharing mechanics; the largest share wins, under 5% means Other.
| Country | Share of visits |
|---|---|
| United States | 56.22% |
English
2 keywords · brand ~46% · organic 100% / paid 0%
| Similarweb top 5 keywords |
|---|
| rpbuddy |
| ai rpg sanbox |
Non-brand queries: ai rpg sanbox
About 38,682 last month (Similarweb estimate), #56 of 606; the month before was 5,333. Opening checkout is not paying, so this is not revenue.
New-user source: Other (Direct / referral / email / paid). Mostly referrals, email or paid, or mostly direct with no clear source Paid social 44.0%, direct 41.4%, search 8.8%; new users mainly come from social ads, and the only search terms are rpbuddy and ai rpg sanbox (brand keywords 45.8%). No breakdown of referral or social sources; returning users come back directly to keep playing their own worlds.
Free tier (50 image credits/month, 50 AI actions/day); $10/month or $100/year; $20/month or $200/year; credit packs $5/50, $8/100, $20/300
Dossier confidence: medium. Parts 1 and 2 come from reading the site and Similarweb data; part 3 from public reporting, founder posts, Wayback snapshots, WHOIS and similar, with speculation marked.
The number: checkout visits is Similarweb's estimate of visits from the site to a checkout payment page; opening checkout is not paying, so this is not revenue. Small sites land on fixed buckets.
Data through last month. Page updated 2026-10-31.