Threads Search Scraper — with engagement metrics

Search Meta Threads by keyword and get back structured posts with engagement already computed, plus a summary that tells you whether the conversation is driven by a few viral posts or spread evenly.

Most Threads scrapers hand you a raw dump of posts and leave you to work out what actually landed. This one ranks and quantifies it for you.

What you get per post

Field	Example
username, isVerified	underthedesknews, true
text	plain-text post body, HTML stripped
likeCount, replyCount, repostCount	1273, 12, 28
totalEngagement, weightedScore	1313, 1449
replyRatio, repostRatio	0.009, 0.021
postedAt, language	ISO 8601 timestamp, en
url, postId, code	direct link to the post

weightedScore counts replies 3× and reposts 5× against a like, because they cost the reader more effort — it surfaces posts that actually moved people, not just ones that scrolled past a lot of thumbs.

replyRatio is useful on its own: a high value means the post started an argument rather than earning agreement.

Summary (key-value store, ENGAGEMENT_SUMMARY)

• totalPosts, totalEngagement, medianEngagement, maxEngagement
• topShare — what fraction of all engagement the top 10% of posts captured. Close to 1 means a couple of viral posts dominate the topic; low means broad participation.
• verifiedShare — how much of the conversation comes from verified accounts
• languages — language breakdown of the results

Input

{
  "queries": ["ai", "machine learning"],
  "serpType": "top",
  "maxItemsPerQuery": 50
}

Each keyword is searched separately and tagged in the output via searchQuery.

Scope and limits — please read

Threads returns roughly 20 posts per search without infinite scrolling, and this Actor does not scroll. Setting maxItemsPerQuery above that will not produce more posts for a single keyword. To collect a larger sample, pass several related keywords rather than raising the limit.

This is a deliberate design choice: it reads only what Threads serves publicly on the search page, without logging in and without driving a browser session. That keeps runs fast and cheap, and it keeps the Actor within what the site returns to any anonymous visitor.

It reads public search results only. It does not log in, does not access profiles that require authentication, and does not target private accounts.

Display names are deliberately not returned. You get the @handle and whether the account is verified, but not the profile's real-name field. Engagement analysis does not need it, and leaving it out keeps the output free of a field that is often a person's actual name. If you need to attribute a post, the url takes you to the source.

Pricing

Pay per result — you are charged only for posts actually delivered to your dataset. Duplicate posts (Threads sometimes repeats results) are removed before charging, so you never pay twice for the same post.

Disclaimer

This is an unofficial tool and is not affiliated with, endorsed by, or sponsored by Meta or Threads. "Threads" and "Meta" are trademarks of their respective owners and are used here only to describe what this Actor reads. You are responsible for ensuring your use of the collected data complies with applicable law and with Meta's terms.