1
2 Comments

Real user, real backend, one honest complaint, one fix

Third build in public update from me, and this one came straight from a user, not from us.

Someone deployed a real backend on livemy this week. An API sitting behind an Android app he ships next week. The deploy itself went great. He told me it saved him hours and a pile of Claude tokens compared to the AWS setup he had been fighting. That part made my day.

Then he opened the API health URL and got our "sleeping" page. On the free tier, idle projects used to go to sleep to save resources. His API was actually fine, the AI confirmed it healthy a second later, but the first thing he saw was a page that looked broken. For someone wiring up a real backend, that is the worst possible first impression.

So we looked at it honestly. A free tier that sleeps made sense for our costs. It made zero sense for the people we actually want: someone building an always on agent, a webhook, a backend for a real app. An app that falls asleep is not live. And our whole pitch is that one word, live.

So this week we changed it. Free tier projects now stay online permanently. No sleeping when idle, even on free. You can build an agent or a backend that is genuinely always up without paying anything to find out whether the idea even works.

If you have something you built with AI sitting on your laptop, this is a decent week to see where it actually runs. Build something that needs to stay awake and tell me where it breaks. That feedback loop is the entire reason the free tier changed.

One line from his email still sits with me though. He said he trusts AWS will always be there, and he does not have that reassurance from us yet. Fair. For an early product, what actually earns that kind of trust for you? Uptime history, open numbers, a real company behind it, something else?

posted to Icon for group Building in Public
Building in Public
on June 10, 2026
  1. 1

    That “one honest complaint” loop is underrated for Android launches. For Kinetic Override, the same pattern applies on-device: if a user hesitates around permissions, timing, or saved profiles, that is not support noise — it is product copy and onboarding data.

    1. 1

      Exactly, hesitation is data, not noise. The tricky part is catching it. On device you can instrument the pause directly, you see the user hover over the permission dialog. On our side it shows up later and messier, as a confused message or a silent churn, so we have to reconstruct the hesitation after the fact. Half the work is just making the signal visible in the first place. Sounds like Kinetic Override has the better seat for that, you get to watch it happen live.

Trending on Indie Hackers
I got my first $159 in sales after realizing I was building in silence User Avatar 53 comments Three Days Before Launch, I Let My Own Tool Tear Me Apart User Avatar 37 comments I thought I was building a news visualization tool. Users thought it was a catch-up tool. User Avatar 32 comments I got tired of rewriting the same content for 9 different platforms. So I built Repostify. User Avatar 30 comments A pattern I keep seeing in EdTech: traffic isn't usually the problem. User Avatar 23 comments I Rejected a $15K Acquisition Offer for My Multi-Agent IDE — Here's the Full Breakdown User Avatar 19 comments