The 12-App Year

I shipped twelve apps last year. Most of them made no money.

That's not a humble brag. It's the point.

I'm a yoga teacher. I learned to build by building. I didn't wait until I could code. I started in Lovable, then Claude Code, and I shipped things. The first one was ugly. The second one was less ugly. By the twelfth one I was making things people actually wanted to use.

Here's what I built:

• Hanuman Chalisa — a devotional app for my yoga students so they could chant without fumbling through PDFs. The only one that makes money.
• Between Us — 75 Valentine's Day questions for couples. Built in a weekend. Used by maybe 200 people for one week. Worth every hour.
• StudioText — SMS for yoga studios, because watching studio owners struggle with email platforms made me want to throw my laptop.
• Ask the Yogi — an AI chatbot trained on yoga philosophy and poses. My students ask it things at 2am that I can't answer.
• Anew — retreat pages written by AI from a three-minute voice note. Built because retreat leaders spend weeks on website copy and it's killing them.
• The Ceiling — a direct-to-fan platform I built as a gentle pitch for an independent musician. Never launched publicly.
• Bloom — a financial wellness app I built as a pitch concept, because finance apps shouldn't feel like Wall Street.
• Pickle Athletics — a strategy deck for Oakland's pickleball social club. Built on a whim after visiting.
• Begin Again Retreat Center — a digital sanctuary for a real one.
• Marin Yoga Festival — a two-day yoga gathering I'm producing.
• Bhakti Flow Church — my monthly Zoom class, with a countdown timer and testimonials.
• Vesper — a private space for families walking hard roads together.

Not all of them worked. Some were gifts. Some were pitches. A couple were experiments that stayed experiments.

Here's what I learned.

Velocity compounds

Each build sharpened my taste. Not my code. My taste. What to leave out. Where to stop. When a feature is making the product worse. You can't learn that from reading. You learn it by shipping something, watching a person try to use it, and realizing you got the tone wrong.

Most apps don't need to scale. They need to land.

Between Us had a shelf life of two weeks. I built it for one person. I'd do it again.

AI doesn't build for you. It builds with you.

If you don't have taste, you'll generate slop at 10x speed. If you do, you'll find your own voice faster than you ever thought possible. The tool rewards the builder who already has a clear eye.

Finishing is the skill

I used to start things and abandon them. My laptop was a graveyard of 60%-done ideas. Now I ship on the first weekend or not at all. Shipping means a working URL I can send to someone. Not a pitch deck about what it could be.

Most of the value is hidden

No one sees the twelve-app year. They see the apps. They see the website. They don't see the compounding. How the Bhakti Flow Church page taught me something I used on Bloom, which taught me something I used on Anew, which taught me something I used on this site you're reading right now.

---

Next year I'll ship another twelve. Probably more. Some will flop. A couple might work. The building itself is the point.

If you're waiting to be a builder until you know how to code, you're waiting for something that isn't coming. Open Lovable. Describe something you want. Click generate. Keep going until it's alive.

That's the whole trick.