LinkedIn Post Search Scraper with AI Relevance Filtering

Find public LinkedIn posts about the topics, brands and people that matter to your business. Turn relevant conversations into a clean spreadsheet with post text, author details, publication dates and available engagement data — without a LinkedIn login or coding.

Enter your topics, choose a market, and click Start. AI checks the meaning of each retrieved post. Keyword search results are saved only when the relevance decision is true.

Why use this LinkedIn scraper?

• Spend less time removing noise. Semantic filtering distinguishes substantive discussion from incidental keyword mentions.
• Start collecting sooner. Built-in parallel search avoids waiting for separate search Actors to start and finish.
• Keep searches bounded. Search effort adjusts automatically to your requested result count. No advanced configuration needed.
• Get useful context. Export post text, author links, reactions, comments, hashtags, images and publication timestamps when publicly available.
• Work without account setup. No LinkedIn cookies, login or personal AI key is needed to use this hosted Actor.
• Avoid duplicate rows. Each LinkedIn post appears once per run, even if found in several searches.

Get started in three steps

1. Add one or more Keywords: a topic, company, person, event or research phrase.
2. Choose the Target country and Max posts per keyword. Optionally set publication dates.
3. Click Start, open Output, and export your results as CSV, Excel or JSON.

Start with 10–100 posts per topic. Requested counts are upper limits: the Actor may return fewer posts when few public candidates are discoverable, LinkedIn blocks access, or AI rejects unrelated content.

Popular use cases

Goal	Example topic	What to look for
Brand monitoring	Enilive	Relevant brand conversations and announcements
Market research	enterprise AI adoption	Public experiences, projects and market signals
Sales research	warehouse automation	People discussing needs and projects in your market
Competitor research	a competitor's product name	Product launches, feedback and positioning
Event monitoring	an industry conference name	Public recaps and attendee perspectives
Content research	circular economy	Themes and examples for your next campaign

Search settings

Setting	What it does
Keywords	One topic per line, up to 100 topics. AI evaluates each post against the topic.
Target country	Tunes search market and language. Defaults to us; try it, gb, de or fr. This is not proof of the author's location.
Max posts per keyword	Up to 100 by default; accepts 10–3,000. Actual availability may be lower.
Posted after / before	Inclusive calendar dates in UTC, also checked against each post's activity timestamp.
Post URLs	Retrieve specific public posts without search or AI relevance filtering.

Search first covers the requested date interval, then explores additional date windows and result pages automatically. Results are not guaranteed to be ordered by date. Sort the exported postedAt column when needed.

A quoted phrase narrows the search query. AI still evaluates semantic relevance rather than requiring a literal match in the post. Public search indexing limits what can be discovered; this is not an exhaustive export of LinkedIn.

Input examples

Find posts about a brand

{
  "keywords": ["Enilive"],
  "country": "it",
  "maxPostsPerKeyword": 20
}

Research several topics within a date range

{
  "keywords": ["enterprise AI adoption", "warehouse automation"],
  "country": "us",
  "maxPostsPerKeyword": 100,
  "publishedAfter": "2026-09-01",
  "publishedBefore": "2026-09-23"
}

Retrieve known posts

{
  "postUrls": [
    "https://www.linkedin.com/feed/update/urn:li:activity:7471837778893459457"
  ]
}

The URL is an illustrative example; replace it with an accessible post. Direct URLs are explicitly requested content: their AI relevance fields are null. When mixing URLs and keywords, direct URLs are processed first and duplicate posts are not saved again.

Output example

Illustrative search result, abbreviated for readability:

{
  "postUrl": "https://www.linkedin.com/feed/update/urn:li:activity:7471837778893459457",
  "postId": "7471837778893459457",
  "text": "Our team is piloting warehouse automation to improve picking accuracy.",
  "authorName": "Example Author",
  "authorUrl": "https://www.linkedin.com/in/example-author",
  "reactionsCount": 42,
  "commentsCount": 5,
  "postedAt": "2026-06-14T08:15:46.000Z",
  "searchKeyword": "warehouse automation",
  "searchCountry": "us",
  "isRelevant": true,
  "relevanceProbability": 0.97,
  "scrapedAt": "2026-09-23T10:00:00.000Z"
}

The dataset also includes author type, headline and image, post URN, relative publication time, edited status, images, hashtags, mentions, video presence and locale when available. Publication time is derived from the LinkedIn activity ID. Engagement values reflect the public embed and may be incomplete; missing counters currently appear as zero. Private profiles and unavailable post text are not included.

How AI relevance works

The Actor checks whether each public post meaningfully discusses your topic. It accounts for synonyms and equivalent meanings and filters out ambiguous entities and incidental mentions. Only positive relevance decisions enter keyword search output.

relevanceProbability is an estimated relevance score from 0 to 1, not a guarantee of correctness. Automated filtering can make mistakes. Public post text and search topics may be processed by external services to provide this feature.

If the AI service fails or returns an invalid decision, the run stops with an error rather than saving unverified results. Already verified results remain available. Direct URL mode does not use the AI service.

Speed, cost and coverage

Parallel processing and duplicate removal reduce repeated work. Search effort is managed automatically, so you only choose your topics, market, result count and optional dates. Actual speed and cost vary with availability and the proportion of relevant posts.

The Run summary shows saved, rejected and unavailable posts and elapsed time. Check the Store pricing panel for current pricing. On an Apify free plan, a limit of 10 output posts per run applies; paid plans remain subject to their configured result cap.

Frequently asked questions

Does this scrape private LinkedIn content? No. It works with publicly indexed URLs and publicly accessible post embeds.

Can I use it without coding? Yes. Fill in the form, run the Actor and download a spreadsheet. No personal API key is required for the hosted Actor.

Why did I receive fewer posts than requested? Some posts are not indexed, are unavailable, fall outside your dates or fail the relevance check. Try a broader topic or wider dates to improve coverage.

Can I automate recurring monitoring? Yes. Use Apify schedules and connect datasets to your existing workflows through Apify integrations or API. Each run deduplicates its own output.

Are results guaranteed to be relevant? No classifier is perfect. Every saved search result received a positive AI decision, and the probability is included for review.

Is every post downloaded again for every keyword? A bounded in-run cache shares downloaded content between topics, while each topic receives its own relevance decision when needed.

Discover more tools for your workflow

Build your research and data collection workflow with more ready-to-run tools. Discover all Lofomachines Actors on Apify.