Bluesky Scraper

Scrape Bluesky through its public AT Protocol XRPC API — no login, no API key, no anti-bot. Two modes:

• Search — pass a searchQuery keyword and get matching public posts.
• Authors — pass authorHandles (e.g. bsky.app) and get each author's profile (followers, bio, post count) plus their recent posts.

Input

Field	Type	Description
searchQuery	string	Keyword(s) to search posts for. Use this or authorHandles.
authorHandles	array of strings	Handles to scrape (the leading @ is optional). For each, returns the profile + recent posts.
maxItems	integer	Max posts per query / per author (default 100). Follows the API cursor until reached.
proxyConfiguration	object	Optional. The Bluesky public API has no anti-bot and needs no proxy, so this is off by default. Only enable it if you hit IP rate limits.

Provide at least one of searchQuery or authorHandles.

Output

Each post row:

{
  "ok": true,
  "type": "post",
  "uri": "at://did:plc:.../app.bsky.feed.post/3kxyz...",
  "postUrl": "https://bsky.app/profile/bsky.app/post/3kxyz...",
  "authorHandle": "bsky.app",
  "authorName": "Bluesky",
  "authorDid": "did:plc:...",
  "text": "…",
  "createdAt": "2024-01-01T00:00:00.000Z",
  "likeCount": 0,
  "repostCount": 0,
  "replyCount": 0,
  "quoteCount": 0,
  "langs": ["en"]
}

In author mode a profile row (type: "profile") is also emitted per handle, with did, handle, displayName, description, followersCount, followsCount, postsCount, avatar, banner, createdAt, and profileUrl.

Posts are deduplicated by uri. The rkey used in postUrl is the last path segment of the post uri.

Nullable fields. Some fields can be null when the API omits them: on posts, postUrl, authorHandle, authorName, authorDid, and createdAt (counts default to 0, text to "", langs to []); on profiles, handle, displayName, description, avatar, banner, createdAt, and profileUrl (counts default to 0).

Diagnostics

If the run fails or returns nothing, a single ok:false row is pushed with an errorCode (BAD_INPUT, NO_RESULTS, RATE_LIMITED, SERVER_ERROR, NETWORK, …) and a human-readable error message. Diagnostic rows are never charged.

Troubleshooting. If you get a BAD_INPUT row, set searchQuery to a keyword or add at least one handle to authorHandles. A NO_RESULTS row means the API answered but had nothing for that query/author — Bluesky's public index is smaller and sparser than Twitter/X, so broad keywords may return few posts. If you see RATE_LIMITED from many parallel runs, enable the optional proxy or lower the volume.

Billing

Charged per unique post returned (post event). Profile rows and diagnostic rows are not charged.

API

Built on the public host https://api.bsky.app:

• app.bsky.feed.searchPosts
• app.bsky.feed.getAuthorFeed
• app.bsky.actor.getProfile

All are public, cursor-paginated GET/JSON endpoints. We hit api.bsky.app directly rather than the documented public.api.bsky.app alias: the alias is fronted by BunnyCDN, which intermittently returns 403 for searchPosts in some regions, whereas api.bsky.app is the same public AppView served directly and is more reliable.