LinkedIn Post Search Scraper

Type keywords, get public LinkedIn posts back as rows: who wrote it, the full post text, the exact timestamp it was published, how many reactions and comments it has, and the link to the post.

No LinkedIn account. No cookies. No session token to paste in. No proxy of your own to buy.

Price

$0.90 per 1,000 posts ($0.0009 per post), plus $0.003 to start a run.

That per-post rate is the lowest on the Apify store for LinkedIn post search. It is flat: the same $0.0009 per post on the free plan as on any paid plan, with no volume tiers, no minimum spend and no monthly fee.

Posts returned	What you pay
10	$0.012
100	$0.093
1,000	$0.903
10,000	$9.003

What actually bills

• $0.0009 per post row. Charged once, immediately after a real post is written to the dataset. The event is called post-found.
• $0.003 per started run. One charge per run at the default 1 GB of memory. The start fee is billed per gigabyte, so if you raise the memory in the run options you raise the start fee with it. There is no reason to: this Actor is built for 1 GB.
• Free: the sample row. Start a run with empty input and you get one labelled row showing the output shape, marked "_sample": true, "charged": false. Nothing is billed for it.
• Free: diagnostic rows. If a query finds nothing, or an input field cannot be honoured, the Actor writes a row explaining why, marked "_diagnostic": true, "charged": false, and finishes successfully. Those rows are never charged.
• Free: everything the Actor fetched and then discarded. A run with a date, author or media filter reads more post pages than it returns, and a search engine sometimes answers with posts that have nothing to do with your keywords. You are billed for the rows you keep, never for the work it took to find them.

There is no free trial run. The list above is exactly what is charged, and nothing else is.

Input

Field	What it does
searchQueries	Keywords to search public posts with. Several queries per run is fine, up to 20.
maxPosts	Posts per query. 0 means "as many as discovery finds", bounded by resultsLimit.
resultsLimit	Hard cap on charged rows for the whole run. Max 1000.
sortBy	relevance keeps search ranking order, date puts the newest first.
postedLimit	any, 1h, 24h, week, month, 3months, 6months, year. Checked against each post's real publish timestamp.
postedLimitDate	Keep only posts published on or after this date. Accepts 2026-01-31, a full ISO string, or a millisecond timestamp.
authorUrls	Profile or company URLs. Narrows results to posts written by those accounts. Works with or without keywords.
contentType	all, videos, images, documents.
includeAuthorDetails	Adds profile URL, avatar URL and follower count. Free, no extra request.
requestTimeoutSecs	Per-request timeout. Default 25.

{
  "searchQueries": ["b2b sales", "hiring software engineer"],
  "maxPosts": 25,
  "resultsLimit": 50,
  "sortBy": "date",
  "postedLimit": "month",
  "includeAuthorDetails": true
}

Run it with empty input and nothing is charged — you get one labelled sample row that shows every field, so you can wire up your pipeline before spending anything.

Searching one person's or one company's posts

Put the profile or company URL in authorUrls and leave searchQueries empty to get that account's recent public posts, or fill in both to get only their posts about a topic:

{
  "authorUrls": ["https://www.linkedin.com/in/williamhgates"],
  "searchQueries": ["climate"],
  "maxPosts": 20
}

Output

One row per post. This is a real row, trimmed only in the text field:

{
  "ok": true,
  "_sample": false,
  "charged": true,
  "postUrl": "https://www.linkedin.com/posts/coquinn_b2bsales-saleswisdom-thoughtleadership-activity-7378846497251106816-cw7k",
  "activityId": "7378846497251106816",
  "authorName": "Corey Quinn",
  "authorType": "Person",
  "text": "What B2B Sales Taught Me About B2B Sales:\n\nI've been in B2B sales for literally years. Here's what I learned...",
  "headline": "B2B Sales: What I Learned",
  "postedAt": "2025-09-30T17:41:37.263Z",
  "reactionsCount": 151,
  "commentsCount": 41,
  "postType": "SocialMediaPosting",
  "mediaType": "text",
  "videoUrl": null,
  "imageUrl": null,
  "query": "b2b sales",
  "retrievedAt": "2026-08-15T13:12:44.907Z",
  "authorUrl": "https://www.linkedin.com/in/coquinn",
  "authorImageUrl": "https://media.licdn.com/dms/image/v2/D5603AQFJV5OVNEHgbQ/profile-displayphoto-shrink_400_400/0/1711075880570",
  "authorFollowers": null
}

Field	Notes
postUrl	The canonical post URL. A repost resolves to the original post here.
activityId	LinkedIn's numeric id for the post. Stable, good as a primary key for deduplication.
authorName, authorType	authorType is Person or Organization.
text	The full public post body, newlines and emoji intact.
headline	LinkedIn's own short title for the post. Often the first line.
postedAt	Exact ISO timestamp, to the millisecond. Not "3d ago".
reactionsCount, commentsCount	The public counters at the moment the post was read.
mediaType	text, images, videos or documents.
videoUrl, imageUrl	Direct media URLs when the post carries them, otherwise null.
query	Which of your searchQueries produced this row.
authorUrl, authorImageUrl, authorFollowers	Only when includeAuthorDetails is true. authorFollowers is frequently null — LinkedIn does not publish it on every post page.

Rows that are not posts are clearly labelled and never billed: _sample: true for the empty-input example row, _diagnostic: true with an errorCode and a plain-English details string for anything that went wrong.

How it works, briefly

LinkedIn's own content search redirects a logged-out visitor to the login page, so this Actor does not use it. What it uses is the fact that every individual linkedin.com/posts/... page is public and indexed: it finds the post URLs through public web indexes, then reads each post page directly and pulls the structured record LinkedIn publishes on it. That record is where the author, the exact timestamp, the body text and the reaction and comment counters come from — they are not guessed from a search snippet.

Two consequences worth knowing about. First, the date on every row is exact, and the date filter is applied before a post page is ever downloaded, so a "past week" run does not waste your run time opening four-year-old posts. Second, if a search engine cannot match your keywords it will sometimes answer with real but unrelated LinkedIn posts. The Actor checks that each post actually mentions something you searched for and drops the ones that do not, uncharged, rather than billing you for them.

Requests go out over rotating datacenter addresses. There is no browser anywhere in this Actor, which is a large part of why the price is what it is.

Limits, honestly

• About 90 posts per query is the ceiling. The public indexes hand back roughly that many distinct post URLs for a given phrase, and asking for more does not produce more. If you need thousands of posts on a topic, split it into several narrower queries — maxPosts applies per query, resultsLimit caps the run.
• Very fresh posts take time to appear. A post published in the last hour is often not indexed yet. postedLimit: "1h" is supported and filtered exactly, but expect thin results; 24h and week are much more productive.
• Coverage is what the public web has indexed, not LinkedIn's internal search. Well-discussed topics, hashtags and named people are covered well. Very obscure phrases are not.
• Some posts cannot be read. Members-only updates, deleted posts and a small number of pages return no public record. Those are skipped silently and never billed.
• contentType values jobs, live_videos and collaborative_articles are not detectable on a public post page. They are accepted so an existing input file still runs, and the Actor writes a free diagnostic row saying the filter was not applied instead of pretending it was.
• authorFollowers is often null. LinkedIn only publishes it on some post pages.
• Counts are a snapshot. reactionsCount and commentsCount are the numbers at the moment of reading, not a running total.
• Reactions are a total, not a breakdown. You get how many, not how many of each type, and not who reacted.
• Comment bodies are not returned. You get commentsCount. The comments themselves are a different job.
• The author filter is applied to the post's own identifier, which is reliable, but a company that posts under several pages needs each page URL listed.
• A run has a page budget, and a very selective filter can hit it. Each run starts with an allowance of pages it may read and earns more as it delivers posts, so a request that needs a hundred pages to produce three matches stops early rather than burning your run time. It writes a free PAGE_BUDGET_REACHED row when that happens, and you are charged only for the posts you got. In practice only heavy contentType or multi-author runs reach it; a normal keyword run does not.
• If every public index is rate-limiting at that moment, the run finishes with a free SEARCH_ENGINES_REFUSED row and no charge beyond the start fee. Running again a few minutes later normally works — the Actor reads three independent indexes and only one has to be up.

FAQ

Do I need a LinkedIn account, cookies or a session token? No. Nothing is logged in and nothing is stored. Only pages LinkedIn serves publicly are read.

Is this legal / is it scraping private data? It reads only what LinkedIn publishes to anyone without an account, the same pages a search engine indexes. No private profiles, no connections, no email addresses, no member-only content.

What is the cheapest way to run it? Ask for what you need. Charging is per returned post, so maxPosts: 20 costs $0.018 plus the $0.003 start fee. There is no penalty for many small runs beyond the start fee, and no discount for one big one.

What happens if my keywords return nothing? The run finishes as SUCCEEDED with one free diagnostic row explaining why, and you are charged only the start fee. A failed or empty run never bills for results. If the first three queries in a run all come back empty, the Actor stops early and says so rather than spending the rest of your run time on the same outcome.

Can I get posts from a specific person or company? Yes — authorUrls. Leave searchQueries empty for everything recent from them, or combine both to filter by topic.

Can I filter by date? Yes, two ways, and they stack. postedLimit is a rolling window (24h, week, month and so on) and postedLimitDate is a fixed date. Both are checked against the post's real publish timestamp, not against a relative "2w" label.

How do I get more than 90 posts on a topic? Use several queries in one run. ["b2b sales", "b2b sales prospecting", "b2b sales pipeline"] will find substantially more distinct posts than any single phrase, because each phrase reaches a different slice of the index.

Does it work in languages other than English? Yes. The post text is returned exactly as published, in any script, and non-English keywords work.

Can I schedule it? Yes, like any Actor. sortBy: "date" with postedLimit: "24h" is a good daily monitor. Use activityId to deduplicate against what you already stored.

What do I do about duplicates across runs? Key on activityId. It is LinkedIn's own id for the post and it does not change.

Can I use my own proxies? The Actor manages its own egress and does not need any input from you. It never routes traffic through metered per-gigabyte proxies, which is what keeps the per-post price this low.

Will it ever fail a run? Only on a genuine crash. Blocks, empty results and bad input all finish as SUCCEEDED with a free diagnostic row, because a failed run still costs you a start fee and being told "you forgot a field" is not worth paying for.

Integration

The dataset is available as JSON, CSV, Excel, XML and RSS through the standard Apify dataset API, so it drops into Make, Zapier, n8n, Google Sheets or a plain HTTP call. To use it from the API:

POST https://api.apify.com/v2/acts/dami_studio~linkedin-post-search-scraper/run-sync-get-dataset-items?token=<YOUR_TOKEN>
Content-Type: application/json
{"searchQueries":["b2b sales"],"maxPosts":25}

That returns the rows directly in the response body.