0
Show Lemmy: I built a free streaming site from scratch — vanilla JS, Node.js, MongoDB, $0/month stack
(eli6movies.vercel.app)
Built ELI6 Movies as a side project to push how far a free-tier stack can go.
Tech:
- Frontend: pure HTML/CSS/vanilla JS on Vercel (free)
- Backend: Node.js/Express on Render (free)
- DB: MongoDB Atlas M0 (free)
- Metadata: TMDB API
- Video: third-party embeds with smart server failover
Interesting bits:
- Multi-server failover via postMessage detection
- Episode email notifier
- Tiered TMDB cache
- i18n (EN/IT/RU)
- Android TV spatial navigation
Happy to answer questions about the architecture.
SLOP
Slop, slop, slop everywhere!
This is shite and you should be embarrassed of the pops and malware you’re exposing people to.
Mate it worse then you think.
The repo
https://github.com/EliseyRotar/eli6_movies
Step 3 in the database setup.
Thanks for exposing yourself to that mess.
I didn’t even go to the repo because of the hypothetical potential for infection! 🤣
It’s encouraging to see people that recognize the glaring issues with these sloppy piggies popping up everywhere.
yeah same answer as the other comment — render free tier has no static IPs so you cant restrict by IP at all. strong credentials are the actual protection, not IP allowlisting
You can. Your AI slop just doesn't know how.
you are right, my bad. render publishes fixed CIDR ranges per region, you can get them from the dashboard and allowlist those in atlas instead of 0.0.0.0/0. updating the readme
popups were coming from window.open() inside the embed iframes, removed allow-popups from the sandbox so that cant happen anymore. should be good now
🤣🤣🤣🤣🤣
You even slop coded your reply here complete with an emdash!
Step away from the LLM and try to UNDERSTAND what you’re trying to get others to use. Your app is like an open sewar and you’re completely blind to what it actually does because you’ve become so dependent on the LLM. What happens when Anthropic or OpenAI start charging what it actually costs to run their hardware? Anthropic says that they see $8/day per average user in token usage. You’ll be stuck with spaghetti code and no LLM to unravel the context for you.
I use LLM’s all the time but at least I absolutely insist on keeping my hand on the steering wheel, closely supervising and making sure to fully understand the cognitive context of what the LLM Is producing. If you don’t even know what it does, you should have your stupid LLM walk you through it at the very least. Until then, no one is going to trust your strange ai fueled fever dream.
I’m embarrassed for you.
got me on the emdash lol. fair points though, working on it
“Fair”
I’ve used Claude enough to know that even this reply was generated by it. Seriously. Wake up!
believe what you want
popups came from the embed iframes, not the site itself. removed allow-popups from the sandbox so they cant do that anymore