Example :

gelbooru.com/index.php?page=post&s=view&id=3919574

On gelbooru, the source is listed as i.pximg.net/img-original/.../19/45/15/60117958_p0.png which just gives me a 403 forbidden error and isn't the correct source.

Same image on danbooru : gelbooru.com/index.php?page=post&s=view&id=3919574

Source is listed as www.pixiv.net/member_illu...um&illust_id=60117958 which is a valid link.

At some point the danbooru bot appears to have started using pixiv links that just gives 403 errors.

We copy the source link on Danbooru's API. The link you're seeing on Danbooru's post pages is generated on the fly and is not visible to our bot, since we do not load post pages.

But if you look at the old posts by the danbooru bot, they dont have this problem and can copy the link just fine? Its not very helpful to have a bunch of 403 forbidden links as the source.

Our script modifies the old format to a working URL, however, with the recent change Pixiv made, or maybe Danbooru made, we have not yet made the code to modify that on the fly so URLs link to the artist's post page.

If you look at those old posts on our API, they'll have the direct image link as well.