Hacker News new | ask | show | jobs
by dom96 1109 days ago
I've been following this since it erupted and am currently developing something that I hope will either enable the Reddit community to transition to a new home or force Reddit to abandon its paid API plans.

The first step in my plan is to implement a read-only API proxy which does not use Reddit.com's API but instead scrapes the necessary data. This should cover 80% of the API traffic, if it works then third-party apps will be able to transition their apps to this new API. So if Reddit does put their API behind a paywall there will be a way for developers to avoid at least some of the ludicrous costs.

I already have next steps in mind, but they really depend on what Reddit does. I sincerely hope they reassess what they are doing with their API.
2 comments
fluidcruft 1109 days ago
Doesn't reddit's website use an API? It's new.reddit.com seems like some single page app bullshit, right?

The real problem you will run into though is that Reddit will restrict what you can access without login. They already tried to flip the switch once on requiring login to access NSFW content, and had to backtrack at least temporarily. I'm not exactly sure why they had to, but supposedly they have now added some missing feature to their web version that allows upload of NSFW content on desktop, so maybe they're ready to unpause. And the official API will block NSFW so I think that's returning. And once you login, scraping will earn a ban. Reddit is also very over zealous with IP bans.

link
dom96 1109 days ago
It's actually interesting how they do it. The new and old Reddit seems surprisingly similar in how they handle at least the initial page loads.

> And once you login, scraping will earn a ban. Reddit is also very over zealous with IP bans.

Argh, that would suck. Not as bad as losing my Twitter but still. That said, they explicitly say in their API docs "We're happy to have API clients, crawlers, scrapers, and browser extensions". So at one point in time they were happy to allow scrapers.

link
rekoil 1109 days ago
Why not just impersonate the official mobile app and use its API?
link
dom96 1109 days ago
That's another option. Scraping the webpages isn't actually that difficult (though admittedly it is likely to be less reliable).
link