1
3 Comments

Week 3: I Hit My Free Tier Limits at 1am on a Tuesday

I posted a weekly update on LinkedIn.

Nothing fancy. Just what I shipped, what surprised me, what I learned. The same stuff I'd been posting every week for Nori, my AI email assistant.

Before going to bed, I give a quick look at my phone. Firebase was down. Not down. Maxed out. The free tier couldn't handle the traffic.

That LinkedIn post went viral. I didn't expect it. I never expect it.

What I learned that week

Lesson 1: Give it away first

I launched a free tier. Listened to Elena Verna on Lenny's podcast saying you earn paid users by giving it away first.

Felt counterintuitive. I'm a solo founder. No VC money. Every dollar counts. And AI apps have variable expenses.

But she was right. Free tier dropped friction to zero. People started trying Nori without needing a pitch. Some stuck around. Some will pay. The data's still early, but the logic is sound. 25 people registered that night.

You can't sell something people haven't tried.

Lesson 2: Friction kills conversion

I had mandatory onboarding calls. Thought I was being helpful. "Let me show you how this works!"

Almost nobody converted.

Killed the calls. Replaced them with a 3-minute video. Conversion jumped.

Turns out people don't want to talk to me. They want to use the product. The call was friction disguised as service.

Conversion went from 0% to 47%.

Lesson 3: Find your enemy

I listened to a branding expert: the best brands have enemies.

If everyone likes you, you're boring.

My enemy? Notifications.

Your inbox shouldn't own your attention. Email should work like physical mail. You check it once or twice a day, not every 30 seconds.

Nori is the assistant that protects your focus, not another thing begging for it.

That framing changed everything. I wasn't just building an AI tool. I was fighting against the notification economy.

People get that. They feel that.

The thing about building in public

I've posted weekly updates for 3 weeks now.

Most weeks? Crickets. 40 views. 10 likes. My mom commenting.

Then one week it hits. 34,000 views. 30+ comments. DMs from people wanting to try Nori.

What I did different: I didn't post an AI-generated preaching post. I posted about my learnings, tagged someone with lots of followers (who liked my post), and attached an image. I think all those things combined made it go viral.

What happened at 1am

Firebase hit the free tier limit.
Nori was down for 3 hours while I scrambled to upgrade to the paid plan. Lost some users.

Also noticed Nori was designed only for desktop, so the onboarding had a dead end for mobile users.

Good problem. Painful problem.

I'm now on the paid Firebase plan. Exploring alternatives for when this happens again. Because it will happen again.

Next week?

Fixing dark mode. The UI is gorgeous on light backgrounds, broken on dark. Also posting more consistently on X.

And probably hitting another limit somewhere. That's how this goes.

See you next week.


I'm Matias, and I'm building Nori: an AI executive assistant that protects your focus. Not another notification. Follow the journey.

posted to Icon for group Building in Public
Building in Public
on January 5, 2026
  1. 2

    Hitting free-tier limits that early is usually a good sign — but it also forces a decision faster than planned.

    At this point, I’ve seen teams benefit most from deciding whether the limit means double down on what’s working or cut scope hard before scaling infra.

    Curious — what’s the one outcome you need clarity on right now to decide your next move?

    1. 1

      I need more data from users and usage in general.

      1. 1

        That makes sense.

        One thing that’s helped me in similar spots is being explicit about which decision the data is meant to unlock — pricing, infra spend, feature focus, or retention risk.

        Otherwise it’s easy to keep collecting signals without actually getting clarity.

Trending on Indie Hackers
I'm a lawyer who launched an AI contract tool on Product Hunt today — here's what building it as a non-technical founder actually felt like User Avatar 142 comments “This contract looked normal - but could cost millions” User Avatar 54 comments 👉 The most expensive contract mistakes don’t feel risky User Avatar 41 comments The indie maker's dilemma: 2 months in, 700 downloads, and I'm stuck User Avatar 39 comments A simple way to keep AI automations from making bad decisions User Avatar 35 comments I spent weeks building a food decision tool instead of something useful User Avatar 28 comments