LinkedIn Post Scraper

Pull public LinkedIn posts into a clean dataset. Paste a post URL, a company URL, a profile URL, or a topic URL — the actor detects the URL type and reads the appropriate posts.

No login required. No account at risk.

What you get per row

• Author name, profile link, whether the author is a person or a company, author follower count
• Post text + a one-line headline
• Posted-at timestamp (ISO 8601, UTC)
• Reactions and comments counts
• Primary image with its accessibility caption (when present)
• Sample of up to ten visible comments — each one a small object with text, authorName, postedAt, and likes
• Hashtags parsed from the post text
• Stable post ID (urn:li:activity:<id> format) for de-duplicating across runs. Long-form articles (/pulse/... URLs) don't have one, so their urn field is null — use postUrl instead for those rows.

Inputs

Paste any mix of these URL types:

URL form	What you get
linkedin.com/posts/<…>-activity-<id>	One row for that post
linkedin.com/feed/update/urn:li:activity:<id>/	Same as above (LinkedIn aliases the two)
linkedin.com/pulse/<slug>	One row for that long-form article
linkedin.com/company/<slug>/	Up to N recent posts visible on the company page (one row each, where N is your "Max posts per company, profile, or topic" setting)
linkedin.com/in/<handle>/	Up to N recent posts and long-form articles from that profile (one row each, where N is your "Max posts per company, profile, or topic" setting)
linkedin.com/top-content/<topic>/	Up to N trending posts on that topic from across LinkedIn (one row each)

Optional knobs:

• Max posts per company, profile, or topic — cap on how many recent posts to read from each of those URL types. The actual count depends on how many posts that source has visible on its public page. A direct post URL returns one row, unless the same post was already emitted earlier in the run (for example, when you also paste the company page that featured it).
• Only posts newer than — skip posts older than this date or window. Pick a specific date (YYYY-MM-DD) or a rolling window like 7 days, 2 weeks, 1 month, 1 year. Leave empty to take everything.

Looking for trending posts on a topic?

Browse linkedin.com/top-content/ to see all available topic pages, pick one, and paste the URL — you'll get up to N trending posts on that topic in one run.

What we don't yet support

LinkedIn shows these surfaces only to logged-in visitors, so we can't read them as a public guest:

• Keyword search-result pages (linkedin.com/search/results/content/…) — paste a topic URL instead for trending posts on a topic (see the Inputs table above)
• Hashtag feeds (linkedin.com/feed/hashtag/…)
• /in/<handle>/recent-activity/all/ profile activity pages
• /company/<slug>/posts/?feedView=all company posts feeds

If you paste one of these URLs, the row's error field will explain that the URL form isn't supported, and the rest of your input continues normally.

Out of scope (so you know what to expect)

These aren't bugs — they're explicit choices about what this actor is and isn't. If you need any of these, this isn't the right tool:

• Authenticated content — we run as a public guest only. Posts that are private, member-only, or behind a follow-gate aren't available.
• Per-reaction-type breakdown (like / celebrate / support split) — only the aggregate count is on the public page.
• Repost / share counts — not in the public payload.
• Video file URLs — videos come back with caption text and a poster image; the video file itself isn't extracted.
• Comment replies / nested threads — commentsPreview returns up to ten top-level comments per post; replies to those comments aren't included.
• Author email or contact info — out of scope for this actor.
• Real-time streaming or webhooks — this is a pull-based actor; use Apify Schedules to refresh on a cadence.
• Cross-run de-duplication — within a single run we de-duplicate by post ID when available, falling back on postUrl for long-form articles (which don't carry a post ID). Across runs, keep your own list using the same precedence — urn first, then postUrl.

FAQ

Do I need a LinkedIn account or cookie? No — we only read pages a logged-out visitor can see, and nothing is stored against your account.

How do I find rows that didn't return a post? Filter the dataset on succeeded:false. Each such row's error field explains the specific reason — sign-in wall, rate limit, post not available, page that doesn't expose its recent posts publicly, etc.

A company, profile, or topic URL came back with no posts — why? Some pages don't surface their recent posts to public visitors (Apple's company page is one example). Some profiles don't expose their recent posts or articles either. Paste individual post URLs from that source instead — those still work.

Why is the comments count higher than the comments I see in the sample? LinkedIn surfaces only a sample of comments inline on the public post page. The total commentsCount is the real number; commentsPreview is the sample.

Why don't I see reactions broken down by type (like / celebrate / support)? Only the aggregate reaction count is exposed on the public post page; the per-type breakdown isn't.

Can I get a video file URL? Video posts come back with the caption text in postText and the poster image in imageUrl. The video file itself isn't available.

What happens when LinkedIn rate-limits the run? Affected rows come back with an error value noting LinkedIn is restricting access right now, so you can retry just those URLs in a later run.

Disclaimer

This actor reads the same public post pages your browser would render without logging in. We do not bypass authentication, do not collect private member data, and do not republish content. Use the resulting dataset in a manner that respects LinkedIn's terms and applicable law.

Built by

apify.com/futurizerush

---

OpenClaw, Claude Code, Gemini, Antigravity, Codex, ChatGPT