Bluesky Post Search Scraper

Find public Bluesky posts by keyword and export their text, engagement, author information and media in a clean dataset. Search is cookieless. Images, videos, quoted posts and link previews are included whenever they are returned with the search results. There are no individual-post detail requests.

What you get

• Post text, publication and indexing timestamps, languages and content labels.
• Likes, reposts, replies, quotes and bookmarks when available.
• Author handle, display name, avatar, account identifier and account creation date.
• Structured images with thumbnails, full-size URLs, alt text and aspect ratios.
• Available video playlists, thumbnails, captions, external previews and quoted-post content.
• Original content metadata including facets, mentions, hashtags, links, reply references and attachment metadata.
• Additional author and post information returned with search results, without an extra enrichment fee.

No media files are downloaded. Media metadata and links stay structured in JSON. CSV and Excel exports may represent these nested values as JSON strings. Missing fields are omitted, not guessed or filled with null placeholders. Zero engagement counts are kept.

Start a search

1. Enter a Search query, such as birds or technology.
2. Choose latest or top.
3. Optionally enter Since and Until as DD-MM-YYYY, for example 07-10-2026.
4. Set Maximum results and start the run.

The actor requests further search pages automatically until it reaches your result limit or available results end. No page-count input is needed. Results are deduplicated within each run. Old maxPages values are ignored.

{
  "query": "birds",
  "sort": "latest",
  "since": "01-10-2026",
  "until": "07-10-2026",
  "maxItems": 250
}

Inputs

Input	Meaning
Search query	Keyword or search expression.
Sort	latest (default) or top.
Since (DD-MM-YYYY)	Optional first day, inclusive, in UTC.
Until (DD-MM-YYYY)	Optional last day, inclusive, in UTC.
Maximum results	Maximum unique posts to save. Default 100; allowed range 1 to 100,000.

Dates are text fields to preserve day-month-year ordering. Leave either date blank to remove that bound. Impossible dates, a missing query, reversed date ranges and invalid limits end with a clear input message and no results. Try birds, 07-10-2026 and 100 as valid examples.

Clean result fields

Field	Contents
text	Full returned post text.
createdAt, indexedAt, scrapedAt	Publication, indexing and collection times.
authorHandle, authorName, authorAvatar	Readable author details.
authorId, authorCreatedAt	Stable author identifier and account creation date.
likes, reposts, replies, quotes, bookmarks	Returned engagement counts.
languages, labels	Language and moderation metadata.
media	Complete rendered media: image URLs, videos, external cards and quoted posts.
contentDetails	Other original content fields, including facets, attachments, reply references and record type.
authorDetails	Other author fields such as associated account information and author labels.
postDetails	Additional post fields not already included elsewhere.
viewer, threadGate	Viewer metadata and reply permissions, when available.
postUri, contentId	Unique post and content-version identifiers.

The extra-information objects preserve new fields without creating duplicate raw-post columns. Original attachment metadata and rendered media are both retained because they contain different information. Post identifiers and media links are content, not internal request addresses. Internal retrieval URLs are not written to results, summaries or logs.

Progress and run summary

Each accepted unique post is saved immediately. Short progress lines show results increasing page by page. The run summary reports saved posts, completed pages, request count, available-hit estimate when returned, result-limit status and a plain-language outcome.

An empty search is a successful run with an explanation. If a later page becomes unavailable or repeats its pagination cursor, already-saved results remain available and the run finishes successfully with a partial-result message. A service outage before any results, or a dataset/storage failure, is an infrastructure error and may fail the run.

Limits

Maximum results is a ceiling, not a promise. Search indexing, visibility restrictions, pagination availability and rate limits can return fewer posts. Some queries allow a first page but restrict later pages; those runs retain their results and explain the partial outcome. The actor does not unlock private, deleted, restricted or paid-only content. No full-history or high-volume guarantee is made.

Only information returned with the search results is included. There are no profile-detail or individual-post lookups, invented metrics, private data or downloaded media files. Reuse must comply with applicable terms, copyright and privacy rules. This actor is independent and is not endorsed by Bluesky.