content
the text behind a link, not the file: send a link as json to
POST /v1/content. reddit answers with the post and its comment tree; x
answers with the post, the post it quotes, and the author's own thread above it.
curl https://api.gronka.dev/v1/content \
-H 'Authorization: Bearer <key>' \
-H 'Content-Type: application/json' \
-d '{"url":"https://www.reddit.com/r/pics/comments/abc123/title/","depth":3}'
options: format is json (default) or text (a
plain-text rendering, comments indented); comments (0 to 20, default 0)
adds that many replies; on reddit, depth (0 to 10) caps how deep they go;
on x, thread: false stops at the linked post; on youtube,
transcript: true (or a language code like "en") adds the
captions as timed lines in post.extra.transcript. the answer is one
Thread: source, url, post,
thread (the author's posts in order), comments (a tree) and
truncated. every item has author, text,
media, links, quoted and stats.
errors use the http status: 404 CONTENT_GONE,
400 UNSUPPORTED_SOURCE, 502 CONTENT_FAILED. sites: reddit, x,
bluesky, instagram, tiktok, danbooru, e621, yande.re, konachan, pinterest, and youtube,
soundcloud and other video sites (title, description, stats and youtube transcripts).