Oksitaine

I took a quick look at AWS to check the prices... I have to say that for serverless services like Lambda, we're not looking at very attractive reductions compared to Vercel where I'm currently...

In the end, I'm not going to waste my time trying to save pennies... However, for other AI services, it's rather interesting because $5/hours for an A100 on Replicate is exorbitant.

Thanks, this website is awesome

I was heading in the direction you just suggested. Before going further, I'm going to reinforce my knowledge of Docker and use Linux OS to set up my servers.

Rushing straight into AWS to manage my infrastructure wouldn't be reasonable, and I risk being less productive for a worse result... Especially since AWS prices aren't that far off for my cloud services for my site compared to Vercel, it's not urgent.

Yes it's true, i mean he LLM costs for Character.ai are 13 times cheaper (thanks to their optimization with KV cache) than if they had simply used the API of an LLM service like Anthropic or ChatGPT.

Insomnia, open-source alternative to Postman for example make HTTP Request like the extension Thunder in VSCode.

I use Insomnia every day.

If you're a bit motivated, I recommend PostgreSQL. I had never used it before, but in two days, I managed to use it correctly on a real project! Here's what I did:

Start with SQL using this video: https://www.youtube.com/watch?v=HXV3zeQKqGY&t=2611s

Then PostgreSQL: https://www.youtube.com/watch?v=qw--VYLpxG4

And for using it in production, here's Supabase and a tutorial to get started with PostgreSQL: https://www.youtube.com/watch?v=dU7GwCOgvNY

That's 10 hours of video, and thanks to that, as a beginner, I managed to create a database for a site with 10,000 users per month. Good luck!

As long as you don't have a SaaS or the costs aren't astronomical, you won't gain anything by changing.

The priority is to have clients, not to optimize your site for 2-3 clients.