REDDIT 原始帖子
Brand new ecom from scratch
How realistic do you think it is to build and run a brand new ecom store from scratch using only Claude + a self-managed VM? I’m getting pretty tired of all the limitations, plugins, and especially the hidden fees with platforms like Woo, Presta, Shopify,…
How realistic do you think it is to build and run a brand new ecom store from scratch using only Claude + a self-managed VM? I’m getting pretty tired of all the limitations, plugins, and especially the hidden fees with platforms like Woo, Presta, Shopify, etc. It feels like you don’t really “own” your store, and everything turns into a paid add-on sooner or later. Has anyone here actually gone the fully selfhosted / AIassisted route?
已收录讨论
been thinking about this too lately and tbh the payment processing alone is gonna be a nightmare without using something like stripe or paypal anyway, so you're still somewhat locked into their ecosystem managing a vm is one thing but dealing with pci compliance, ssl certs, backups, ddos protection - thats a lot of moving parts when you could just focus on actually selling stuff. my mate tried building his own cms for his gaming merch store and ended up going back to shopify after 6 months because he was spending more time fixing server issues than growing there business claude's great for automating the tedious bits though, might be worth starting with a minimal viable setup and seeing how far you get before the complexity hits
Why not... tbh this is clearly where things are heading. Better to adapt early than play catch-up later
Yeah I think so too. Honestly, more than anything I’m just curious if anyone here has actually tried going down that route and can share some real experience. Feels like it’s doable in theory but I’d love to hear how it plays out in practice
Honestly, integrating Stripe took me less than an hour with Claude on a previous project. Paypal is a hard pass for me anyway. SSL (Let’s Encrypt which is free), ddos, backups, etc. are all handled by the control panel, so that side of things hasn’t really been an issue. Has your friend actually tried this recently? I’m asking because when I first tested something similar back in Sept, it was a joke.. tried again in Dec and it was already improving, and about a month ago I got seriously hooked after building an app with Claude that would’ve taken months for a team of full-stack devs. This thing is evolving exponentially, not linearly
Use Shopify and actually focus on trading your business.
Where are things heading, champ?
Agreed. You can sign up for the Shopify Starter plan…. $5/mo. And then use the Storefront API to build your front end, totally unrestricted. If you need more functionality on the back end, use the Admin API to whip up and features you need. Being a developer is handy, but there are some really solid, mature features with A platform like Shopify, that you’d be foolish to waste time rewriting.
I think that really depends on what you’re trying to build. In my case, the goal is to create a fully functional store using Claude and a fast performing cloud environment, so I’ve already integrated Stripe for payments. From what I see, Ecwid mainly helps if you want an outofthebox ecom solution with built-in inventory, admin tools, etc. But if you’re building everything yourself, it doesn’t add much value and can actually limit flexibility
This is so off topic
Late last year I wanted to see where I would get building a basic Django backend shop from scratch and it's pretty doable, especially using something like Django that has piles of docs and best practices already figured out. Lately I've been working on other business tools and haven't revisited what I built but there's no reason you can't build your own shop. This sub is mostly a Shopify fanboy sub so you're going to hear from people that disagree.
I’m starting to wonder if there’s some Shopify affiliation here… Defending an expensive SaaS that’s slower than humble and free platforms like OpenCart feels less like helping and more like promoting. Not that the question was about which ecom platform is a best fit for my project... my original post was about sharing experiences, not pushing platforms
Have you actually tried building something like this yourself? Before even considering this approach, I tried building a project management tool like monday as a personal project only using claude+github. In just one day, I ended up with something even better fully customizable and free. That’s when I realized DIY has a different meaning now.
I actually did some testing on another project and Claude (opus 4.6) was able to set up a stripe account and integrate payments directly in the app which was pretty impressive. So I’m thinking the rest might also be manageable with the right approach. I haven’t run this fully in production yet, so if anyone has reallife experience doing this endtoend, I’d love to hear how it went.
这条评论已被删除。
There's a reason the Walmarts, Best Buys, Under Armors etc. of the world pay those fees. It's better than bringing everything in house.
这条评论已被删除。
Totally doable if you’re comfortable managing servers, but the real work isn’t the store code, it’s payments, security, taxes, and all the boring edge cases. AI can help build stuff faster, but you’ll still end up rebuilding a lot of the “boring” infrastructure platforms already solved.
Dear Lord no.
It’s time for an update, as 21 days have passed since we started this discussion. After all, I decided to give it a try, and over the past two weeks I was able to create a fully functional online store with PayPal and dropshipping integration. My technology stack was: Frontend: Next.js 14 + TypeScript + Tailwind CSS + shadcn/ui Backend: NestJS + TypeScript + Prisma ORM Database: PostgreSQL (Railway managed) Authentication: JWT + bcrypt + Passport.js + httpOnly cookies Payments: PayPal OAuth2 for revenue verification Package Manager: pnpm monorepo (apps/web, apps/api, apps/chain, packages/shared) To reply to some of the most recent comments here: first, please stop promoting Shopify. I understand that some of you might be affiliated with them or simply be fans, but the truth is that Shopify is not only expensive, it also has many functional and structural limitations that start to affect a business once it begins to grow. I’m speaking from personal experience. I wanted to build something myself with AI not only because it gives me the freedom to create whatever my imagination comes up with, but also because it lets me eliminate third-party fees for good. With a single Max Claude license, I’ve already removed around a dozen software subscriptions from my credit card, and this is just the beginning.
It's totally doable. You'll need to pick your card processor and integrate with them. I'll suggest Stripe. Another option is to self build the site and farm out the ecommerce part using Ecwid. It's the right direction, just too many people have drank Shopify's Kool-Aid to get it.
You need to make a brand
Get a custom website, with right people on your side everything can be coded 🫡
I think it's easier to just use Shopify. It is not that expensive. There are some great platforms you can connect to it, like Forstock. And you can use the API too. Just get started. The sooner you do, the sooner you might start profiting.
Unless your main goal is to simply learn how to build it all from scratch this would be insanely stupid.
这条评论已被删除。