I have great news for several reasons. First of all, I have documented the whole devRant API now with some help of AI to resolve requests and made a script to resolve the responses and merged that together in a README.md and updated the docstrings of my devranta api client.

This is a big deal, because with such complete api class, you can vibe whatever advanced devRant bot you can imagine in a hour or so. Could make a wordle bot like Snek has. Just an example.

But besides that, I vibed a devRant API crawler using my devRanta class. It searches what users are connected to what rants and reversed since dr doesn't allow to browse the whole site per pagination. Wonder why.

The crawler is in the examples/crawler directory. If you have python3 already installed, the only think you have to run is go into that folder, and it will download whole devRant.

I hope that my work will contribute to the creator of the next devRant or for the training of an LLM or such. It's a nice dataset. The crawler pulls on my terrible slow 4g connection 272,340 rants per hour. The crawler is an beast actually and easy to fine tune regarding workers etc.

Long story short: get the whole devRant dataset by executing once `make`. And don't worry, the API can handle it easy.

@Lensflare, is it already possible to delete your account in your app? I've documented that as well :P

While a lot is vibed, it required quite some steps to pull all off together. I hope somebody takes my advise to vibe some nice bots with it.

If you take my example project files and the devRanta client yourself, you have enough information to generate a whole devRant site clone easy. I think in one day, you can have like 80% from the default user functionality. Maybe even a good idea to also give a few screenshots to the LLM.

https://retoor.molodetz.nl/retoor/...