Lorelightai Hits $16.4K MRR
Benjamin Houy announced that @copycatcafeapp has reached $16.4K MRR in March, marking a 49% growth since November. The milestone reflects the app's success in improving trial-to-paid conversion rates and enhancing product features like flashcards and language support.
April 1, 2026$16K MRR Milestone
Original Announcement
March recap for @copycatcafeapp:
$16.4K MRR (+49% since Nov)
1,012 paid subscribers
58% trial-to-paid conversion
3.7% monthly churn
315 commits shipped
2 cofounders + Claude as dev partner.
What we built:
Rebuilt flashcards from "flip and self-grade" to "type or say it and AI checks your meaning."
Trial conversion jumped from 36% (Feb) to 58% (March). Not all from flashcards — but making the product stickier moves everything.
Migrated from Puma to Falcon. Response times: 251ms → 124ms. Chat uses SSE streaming (one long connection per user). Puma blocked all other requests during these. Falcon handles them on fibers. Night and day under load.
Spanish launching in April — included in every subscription.The entire codebase is now language-parametric: LLM prompts, TTS, STT, pronunciation scoring, views, share cards. Adding a language is a content problem now, not an engineering one.
I handle product and dev. @emerald_springs handles marketing and PR. Claude (Opus) is our dev partner.315 commits in one month. I still make every product decision — Claude does the implementation, research, and testing.
$16.4K MRR (+49% since Nov)
1,012 paid subscribers
58% trial-to-paid conversion
3.7% monthly churn
315 commits shipped
2 cofounders + Claude as dev partner.
What we built:
Rebuilt flashcards from "flip and self-grade" to "type or say it and AI checks your meaning."
Trial conversion jumped from 36% (Feb) to 58% (March). Not all from flashcards — but making the product stickier moves everything.
Migrated from Puma to Falcon. Response times: 251ms → 124ms. Chat uses SSE streaming (one long connection per user). Puma blocked all other requests during these. Falcon handles them on fibers. Night and day under load.
Spanish launching in April — included in every subscription.The entire codebase is now language-parametric: LLM prompts, TTS, STT, pronunciation scoring, views, share cards. Adding a language is a content problem now, not an engineering one.
I handle product and dev. @emerald_springs handles marketing and PR. Claude (Opus) is our dev partner.315 commits in one month. I still make every product decision — Claude does the implementation, research, and testing.
Founder
Benjamin Houy
@BenjaminHouy
Product
Lorelightai
View Product