The Vietnamese Solo Dev Who Quit His Job to Build Feynman AI — Growing a Study App to 370K Users and ~$19K/mo with No Ad Budget, ‘Volume over Virality’ TikTok, and One Rename

Background & product

Feynman AI is a study app: upload a lecture recording, podcast, PDF or YouTube video, and the AI turns it into organized notes, mind maps, flashcards and quizzes. Each note has a built-in chatbot you can ask about the content or pull key points from. Its namesake — the ‘Feynman Technique,’ the idea that you’ve only understood something when you can explain it in child-simple words — is the core experience, selling *retention* rather than mere *capture*.

It was built by Vietnamese solo dev Alex Nguyen (real name Nguyen Thanh Duy, @alexcooldev on X). As of December 2023 he was a software engineer earning $3,200/month; he quit corporate work to go all-in on B2C mobile apps. He launched in October 2024 as ‘Notewave AI’ and hit $20K/mo in about seven months, then renamed the app to ‘Feynman AI.’ The reasoning was simple: a name that *is* a study method tells users ‘what this app is’ with zero explanation, making both marketing and virality easier.

Growth was relentlessly zero-budget organic. Alex is blunt about it: TikTok is ‘the greatest distribution channel in the world for mobile apps.’ And the aim isn’t a single viral hit but ‘volume over virality.’ He runs many TikTok accounts in parallel, uses AI to brainstorm ideas and draft scripts, and shows up every single day. He doesn’t try to invent hits from scratch — he reverse-engineers formats that already work (fast-paced slideshows, ‘study hack’ videos). Layer App Store SEO (ASO) on top and it pushed the app to ~370K cumulative users.

He also refused to fight on iOS alone. His web version crossed $14.6K/mo, overtaking iOS revenue — and because the web has no 15% Apple cut, margins improve and the customer (their email) stays in his hands. Beyond Feynman AI he also runs the medical app MedDeep AI (~20K users, $6–7K/mo); across the whole portfolio he makes $41K–$47.5K/mo. And he publishes his monthly revenue breakdown on X — building in public is, for him, both an acquisition engine and a discipline mechanism.

The repeatable playbook

1. 1Make the app name *be* a known concept so it carries a story (Notewave → Feynman AI, a study method) — it lands with zero explanation and spreads more easily
2. 2On TikTok, choose volume over a single viral hit: run many accounts in parallel, let AI mass-draft scripts, and show up every day
3. 3Don’t invent hits from scratch — reverse-engineer already-growing formats (fast slideshows, ‘study hacks’) and swap them onto your app
4. 4Don’t fight on iOS alone: ship a web version to escape Apple’s 15% cut and reclaim margin and the customer (their email)
5. 5Publish your monthly revenue breakdown on X so building in public does double duty — growing X as a channel and enforcing discipline
6. 6Once distribution wins you scale, deepen the learning experience (active recall — flashcards, quizzes, the Feynman Technique) so retention/LTV becomes the moat

Deep dive

【Deep dive】We break down how a zero-ad-budget solo dev climbed to ~$19K/mo in the red ocean of study apps. The tech is ordinary; he won on distribution design.

■ Turn the app name into a *story* — Notewave → Feynman AI. The first name, ‘Notewave,’ sounds nice but explains nothing. Alex renamed it ‘Feynman AI.’ The Feynman Technique — ‘understand by explaining hard things simply’ — is a study method most people can infer from the name alone, so the name itself states the product’s promise (use this and you’ll actually understand). In his own words, ‘a better name makes marketing and going viral easier and tells the product’s story to users more naturally.’ Lifting acquisition efficiency by changing *only the name*, without touching the product, is one of the cheapest levers an indie has.

■ TikTok is ‘volume over virality’ — a factory of many accounts × AI scripts. Many builders chase one viral hit and burn out; Alex inverts it — hits are probabilistic, so maximize attempts. He runs many TikTok accounts in parallel, hands ideation and script drafts to AI, and mass-produces posts like a factory. If one account is banned or stalls, the others keep running, spreading platform risk. CAC is near zero, and every daily post doubles as a live market test of which angles land.

■ Don’t invent hits — reverse-engineer formats that already win. The volume isn’t reckless. He studies formats that are already growing (fast-paced slideshows, ‘study hack’ videos), then reproduces their structure (hook → payoff → CTA) swapped onto his app. Reproduction beats creation on hit-rate and repeatability. The indie lesson: translation of winning formats over originality.

■ Use a web version to escape the ‘15% Apple tax’ and reclaim profit and customers. Alex didn’t stay boxed in iOS; he shipped a web version — which reached $14.6K/mo and overtook iOS. Web has no Apple/Google fee (15% even under small-business programs), improving margins, and it lets him reclaim the customer relationship (email, etc.) that the app stores otherwise hold. Even for a mobile-born app, moving some revenue and customer assets *outside* the platform is a move that pays off for long-term survival.

■ Build in public — monthly revenue posts do double duty as marketing *and* discipline. Alex publishes his monthly revenue breakdown (dollars by Feynman AI / MedDeep AI / sponsored posts…) on X. Transparency earns follower trust and engagement (so X, itself an acquisition channel, grows), while the public commitment — ‘next month I’ll update this number’ — functions as self-discipline. Indie work is lonely and easy to coast on; publishing becomes a mechanism you can’t quit.

■ The moat is not the tech — it’s ‘learning experience × a distribution pattern.’ The core tech (audio/PDF → summaries and cards) is imitable on its own — indeed, many clones by *other* developers also call themselves ‘Feynman AI.’ So where is Alex’s moat? (1) Polishing the learning experience that forces active recall (retention = LTV), and (2) establishing a *repeatable distribution pattern* first — mass TikTok + ASO + web — to get ahead on scale and review assets. Building the moat into how he *distributes* and what he *retains*, not into how he *builds*, is the biggest reproducible takeaway for indies.