Ben, a non-technical founder and former content creator with 1M+ followers, launched Canary — a language-learning app built around songs — nine months ago; last month it generated $25K in revenue with 50K downloads and 2K paying subscribers.
The app and business model
Canary lets users learn languages through songs they love: lyrics with translations, tap-to-save vocabulary, quizzes, a speed-translation game, and a community feature that matches learners with native speakers for practice.
Primary users are people who tried Duolingo but found it ineffective or boring and want something more engaging.
Simple subscription model: $10/month or $79/year; 60% of subscribers choose the annual plan.
Current metrics: ~$25K revenue in the last 28 days, ~$14K MRR (lower due to annual-heavy mix), 23K active customers last month.
Origin: from one viral post to validated concept
Ben moved to the US 13 years ago, sold a business three years ago, became a digital nomad and content creator.
A year ago he posted a video of himself singing in French with lyrics and translation; it drove high engagement and follower growth.
He created a dedicated theme page (not tied to his personal brand) for song-based language content; it grew to 200K followers in under a year, proving demand independent of his personal audience.
The insight: followers using Duolingo couldn’t speak French but remembered song lyrics — suggesting music as a stronger memory hook for language.
Building as a non-technical founder
Hired a developer for the first MVP (about one month, “ugly and bad”) with a launch-fast-get-feedback mindset.
Four months ago began “vibe coding” himself using AI tools (Claude), which dramatically increased speed of iteration and feature development.
Lean tech stack: Firebase backend (~$200), RevenueCat for subscriptions/paywall testing (~$260), PostHog for analytics/experimentation (~$250), Claude for coding (~$200), Beehive for email (~$169), OpenAI for in-app AI features (~$150), AppsFlyer for install tracking (~$97), ManyChat for social automation (~$30).
Core routine: every day learn something, ship something, test something — heavily inspired by Starter Story episodes; even paid for consulting from past guests (e.g., Nicole on onboarding/product).
Tested content across all platforms and formats; found Instagram Reels performed best and doubled down there.
Used ManyChat to convert organic engagement into traffic (e.g., DM automation from comments).
Paid acquisition: boost top-performing organic content as Meta ads; test creative, audience, caption; scale winners, kill losers; repeat.
Chose to focus on Instagram and TikTok rather than spreading thin across every platform.
Founder philosophy and advice
“Perfection is the enemy of good” — launch an ugly MVP, get real feedback, improve continuously.
Contrasts with peers who spend years building “perfect” apps that never launch; believes “ready is always a work in progress.”
For idea generation: spend time in social media niches you enjoy; look for complaints and struggles — the best ideas hide in the content you’re already consuming.
