How much time do you actually lose to infra setup before writing a single line of product code?
Every developer I talk to has the same story. The idea is clear, the stack is decided, but before anything real gets built there's 3 hours of AWS configs, Dockerfiles, environment variables, and CI pipelines that have nothing to do with the actual product.
AI coding tools made writing code faster. Nobody has solved what comes after.
Working on something early stage in this space. Not ready to share details yet but want to understand how real this pain is for other builders before going too deep.
How do you currently handle infra setup? And what part of it do you hate the most?
155 views
Replies
The tricky part is abstraction. Builders want less setup work, but they also don’t want to lose flexibility or control once things scale.
Product Hunt Wrapped 2025
@nikita_jain18 that's the concern I thought about most honestly. the approach I'm taking deploys everything into your own accounts, Vercel, Railway, Neon. you keep full access and can manage things directly if you need to. no lock in.
This pain is very real 😅 Sometimes I spend more energy wiring services together than actually building the product idea itself.
Product Hunt Wrapped 2025
@greffin_dony right? the wiring is the part that has nothing to do with your actual idea but somehow eats half the day. that's the exact problem I'm fixing.
Hedy AI
CI/CD setup is where my motivation usually drops 😂 By that point I already want to work on the actual product experience.
Product Hunt Wrapped 2025
@kate_sleeman CI/CD is where the energy just dies lol. you started the day excited about the idea and by the time you're debugging a yaml file you've completely lost the thread.
I think developers tolerate infra pain because it became normalized, not because it's actually a good experience.
Product Hunt Wrapped 2025
@lucy_rolff normalized is such a good way to put it. nobody chose this experience, it just became the default and everyone accepted it.
The pain is real. I've abandoned more projects in the infra setup phase than I'd like to admit. It's not even hard work — it's just tedious work that feels completely disconnected from the thing you're actually trying to build.
Right now I default to Railway or Vercel to skip as much as possible, but even then, the moment you need anything custom, you're back to wrestling with configs and environment variables. The part I hate most is keeping env vars in sync across environments. It's boring, easy to mess up, and somehow breaks things in the most silent ways.
Product Hunt Wrapped 2025
@crystalmei The silent env var failures are the worst kind because you don’t know something’s broken until it matters. Railway and Vercel buy you time but you’re right, the moment you go custom you’re back in the weeds. That gap between ‘works on Vercel’ and ‘works everywhere you need it to’ is exactly what I’m looking at
@alexcloudstar That's fantastic! I will pay attention to your project!
Product Hunt Wrapped 2025
@crystalmei awesome! I made a waitlist as well if you want to join https://www.deploynow.sh
@alexcloudstar Thank you！
We have a dedicated devops person in the team who is doing that, so I don't have to bother - but if I do, it's usually the last thing I do, because the first environment for the code to run in is my laptop, and I don't need any special configuration
Product Hunt Wrapped 2025
@sk_uxpin Laptop-first makes sense until you need to onboard someone or push to prod and realize nothing’s documented. Does your devops person own the whole pipeline or are you still involved in any of it?
@alexcloudstar yeah, well, I guess I was referring to an earlier stage of the product than you did :) The devops team does own the whole pipeline
The context is what kills momentum . Going from product thinking AM permissions YAML debugging completely breaks the creative flow.
Product Hunt Wrapped 2025
@bernard_lewis context switching is the real killer. you're in product mode and suddenly you're reading AWS docs. that's exactly the flow I'm protecting
The worst part is that infra setup is rarely difficult intellectually it’s just fragmented, repetitive, and full of hidden edge cases that break flow before real product work even starts.
Product Hunt Wrapped 2025
@bruce_warren fragmented and repetitive is exactly it. nobody's saying infra is hard, it's just soul-draining to do the same dance every single time. that's what I'm automating
TrackerJam
Most indie hackers I know reuse the same templates over and over because nobody wants to relive the setup phase from scratch🚀
Product Hunt Wrapped 2025
@maklyen_may templates help but you're still maintaining them, still updating them, still missing things when a new project has slightly different needs. it's a workaround not a real fix.
Used to be in that boat before, but lately some creative strategies around leveraging vibe coding as the starting point to get to the solution and then creating a framework around migration (either virtualization or structured lift and shift) has cut down time and effort dramatically, either way its important to know what you're building, what your assistant tools are wiring up. in other words, the tech stack understanding and overview is still needed, but can cut costs and efforts significantly if you take a very planned approach from the start. That's been my experience.
Product Hunt Wrapped 2025
@jthomas2174 The vibe coding to framework approach is interesting what does the migration planning actually look like in practice? Curious if you’re templating the lift and shift or rebuilding it fresh each time
@alexcloudstar we have sort of templated it, the major win is dev/solutioning time vs deployment, but with the new approach we still accomplish significant improvement in time to market. The simplest way to describe our approach is the following vibe code --> Online Git --> Parallel Git (for lift and shift) -- > Code + App logic separation (App Logic/Data wiring with live artificats happen here, code stays untouched). So vibe coding platform changes can keep happening real time, the only net delta catch up needed each time is the live wiring to internal application/data. its takes a bit time to establish the framework that works for you, but once set it and the engineers understand the approach it becomes a breeze.