LinkedIn Buying Intent Leads: AI-qualified posts, pay per lead

Find the people on LinkedIn who are asking for what you sell this week — founders asking for a CRM recommendation, teams switching away from your competitor — with the post as evidence and a reply ready to send.

A LinkedIn post search that returns buyers of a product like yours, not posts — no cookies, no login, and you pay only for the people who qualify.

⚡ First 25 leads free · 💵 $0.002 per qualified lead · 🤖 $0.009 with AI assessment · 🔑 No API key needed on a plan that can run Store Actors (else bring your own key) · 🍪 no cookies, no login · ⏱️ demo in seconds · AI off: the search + < 1 min · AI on: + at most 5 min

Try it in 30 seconds, free. Click Try it — the form is pre-filled with a free demo: 12 sample posts, 5 buyers among them, nothing charged. Then type your phrases and what you sell, set Where do the posts come from? to Search LinkedIn and press Start: 3 phrases × 50 posts ≈ $0.30 on your account (capped by Max source spend), and your first 25 real leads here are free. The search needs an Apify plan that can run Store Actors; otherwise use Existing Apify dataset. → What a run costs · What the AI adds · Run it weekly

⚡ At a glance

What you get	From a LinkedIn post search, ranked authors with buying intent: who they are, what they need, a score, a draft comment + DM
You provide	One sentence on what you sell — or your own phrases, a phrase pack, competitors or target accounts
Output	JSON / CSV / Excel, best first, in Shortlist, Leads, Outreach and Signals views, plus the OUTPUT funnel
Typical run	5 phrases × 50 posts = 250 posts scanned → 5–25 qualified leads (estimate; every run's OUTPUT reports the real numbers)
Run time	the search + under a minute for the rules; the AI pass stops after aiTimeBudgetSecs (5 min); OUTPUT has runSeconds. Set .call() timeouts to ≥ 15 min
Cost of that run	≈ $0.50 source search on your account + 20 × $0.009 here + ≈ $0.12 AI tokens = ≈ $0.80
Free tier	a free demo run on sample posts, then the first 25 qualified leads per account; no start fee on any run
Keys / setup	no API key and no LinkedIn cookies — but the search needs an Apify plan that can run Store Actors (the free plan's $5 credit covers ≈ 15 first searches), and so does Apify's built-in AI model access; on other plans use dataset mode and llmProvider: "byok" with your own model key
Which plan	Free demo and your own list (sourceMode: list): any plan · live search: a plan that runs Store Actors
Works with	Schedules, webhooks, Slack, CRMs, Google Sheets, Make / n8n / Zapier, MCP & AI agents

🎯 What this Actor does: LinkedIn intent signals → qualified leads

LinkedIn buying intent is public: people post "can anyone recommend…", "we're moving away from…", "looking to hire an agency…". This Actor turns LinkedIn post leads into a short list of buyers. For every qualified lead you get:

• Buyers, not vendors — six intent types with the matching phrases quoted verbatim; vendor promotion, job ads and empty reposts are dropped free.
• One lead per person — all of an author's posts fold into one lead: you pay once per person, not per post.
• Who they are, for free — role, company and seniority from the headline; decision-makers score higher.
• Only new people, every run — memory across runs, a 30-day cool-down and a suppression list.
• Qualified against your offer — a rule score or an AI fit score with reasons, a public comment and a DM opener.

Built for social selling at a daily or weekly rhythm; AI agents and n8n / Make / Zapier can call it as a tool.

Shape who you keep

• Phrases from your offer — suggestPhrases (auto): with nothing typed, 8–10 buyer phrases are written from offerDescription. Output: matchedQuery on every lead.
• Competitor watch — competitorNames drops their employees; expandCompetitorPhrases (on) adds "alternative to X" searches. Output: mentionedTools.
• Who counts — targetRoles (e.g. "VP Sales"), targetSeniority, includeCompanyPages (off: company pages are dropped). Output: role, seniority, isDecisionMaker.
• What the post says — mustMentionKeywords (all must appear), excludeKeywords (any drops the post), excludeAuthorKeywords (headline words that drop the author, e.g. "recruiter"). Output: intentCues.
• Cool-down — authorCooldownDays (30): a person you already received is not delivered again for that many days. Output: dedupeKey.

🆚 Why this instead of a LinkedIn post search or LinkedIn post scraper?

A keyword search returns every post with the words — the vendor selling a CRM outnumbers the founder asking for one, and the raw scraper bills both.

	Linkedin Post Search Scraper (harvestapi/linkedin-post-search)	LinkedIn Buying Intent Leads
What you pay for	every post, $0.002 each — vendors, job ads and duplicates included	only qualified leads; every rejected post costs $0 here
Start fee	$0.00005 on every run	none — a daily monitor that finds nothing is free
Buyer vs vendor	no intent classification	intent type + verbatim evidence, 14 free noise rules
Unit	one row per post	one lead per person, other posts in otherPostUrls
Who the author is	profile look-up at $0.002 (main) / $0.004 (full) per author	role, company, seniority from the headline at no cost
Repeat runs	the same viral posts again	cross-run memory, 30-day cool-down, suppression list
What to say	nothing	draft public comment + DM opener, empty for cold leads

💵 Pricing — what a run really costs

Pay per result: one event per qualified lead, nothing else.

Event	When it is charged	Price
demo	the pre-filled free demo on the 12 built-in sample posts	$0.00
free-tier	your first 25 qualified leads on this Actor, in any mode (billedAs shows the tier they will cost afterwards)	$0.00
qualified-lead-basic	AI off — intent cues, role, company, seniority, flags, rule score, template replies	$0.002
qualified-lead-ai	AI on — everything above plus AI intent with evidence, need, urgency, fit score, reasons, comment + DM	$0.009

You are never charged for: the demo, vendor posts, job ads, reposts, company pages, posts that fail your filters, leads below minScore, leads the AI marks as "not intent", duplicates, people in the cool-down window or on your suppression list. There is no start event, and a search that cannot run fails with the reason instead of charging.

Worked example — 5 phrases × 50 posts = 250 posts scanned → 20 qualified leads with AI on:

Line	Where it is billed	Amount
Post search: 250 posts × $0.002 (+ $0.00005 start)	the source Actor, on your account	$0.50
20 qualified leads × $0.009	this Actor	$0.18
AI tokens, ≈ 2,450 per lead (measured) → ≈ $0.006 at Claude Haiku 4.5 prices	Apify model access, at cost	$0.12
Total		≈ $0.80 → $0.040 per qualified lead

The basic lead at $0.002 is below the $0.005 median unit price of the 10 closest Store Actors for "linkedin buying intent" (Store audit, 2026-10-03, 8 of them priced); the AI lead at $0.009 is the premium tier. The same run with AI off: $0.54, and on your first real run the 25 free leads cover the $0.18. OUTPUT.aiTokens holds the measured tokens, OUTPUT.yieldByQuery the search dollars per lead of every phrase.

🚀 How to use it

1. Click Try it, then Start — the pre-filled form is the free demo: 12 sample posts, 5 buyers, nothing charged. The status line then tells you the next step.
2. Type the phrases your buyers write ("recommend a CRM") or pick a ready phrase pack (or leave both empty: phrases are suggested from your offer) — and replace the example offer with what you sell.
3. Set "Where do the posts come from?" to Search LinkedIn, add your competitors (competitor watch searches "alternative to X" for you) and press Start. The log names the cost first.
4. Open the Shortlist view — hot leads first, with the post link and the next step. Export or wire a webhook.

🤖 What the AI tier adds

Every quote is checked against the post and cold leads get no text — five minutes of reading and writing saved per lead.

On any plan: the demo shows sample AI verdicts (labelled aiModel: "sample", nothing charged); on a plan without Apify's model access your first 5 real leads still get live AI verdicts on us — all of your free leads need Apify's model access or your own key (a free Gemini key from aistudio.google.com/apikey works for a few per run).

Free AI preview: with AI off, the top 2 leads of a real search still get the AI assessment at the basic price (aiTeaserLeads, aiTeaser: true).

⚙️ Input

The fields a buyer decides on — all 52 options, with defaults, are in the Input tab:

Field	Default	What it does
intentKeywords / intentPacks	— / []	One post search per phrase; packs add 8 proven buyer phrases each (crm_sales, saas_tool, agency_freelancer, recruiting, buyers_de …)
offerDescription	""	What you sell and to whom — the score and the replies are written against it
competitorNames	[]	Their employees are dropped; competitor watch adds 5 switching searches per competitor
sourceMode	actor — with nothing to search (no phrases / packs / accounts / dataset) the free demo runs, nothing charged	actor searches LinkedIn, dataset reads an existing HarvestAPI posts dataset, list qualifies pasted posts
postedWithin / postsPerKeyword	week / 50	Search window (the first live run of a search reads the last month); posts per phrase ($0.002 each on your account)
watchAuthors / watchCompanies	[]	Restrict the search to these people / companies — account intent monitoring
minScore / maxQualifiedLeads	55 / 50	Below the score = not delivered, not charged (70+ hot); hard cap on output and on what you pay
enableAi / llmProvider	true / apify	AI assessment on/off; Apify model access or your own key (byok + llmApiKey)
maxDiscoveryChargeUsd	2 (the form suggests 0.3)	Hard cap on the search's spend on your account — API and MCP calls without it are capped at $2

Everything else — filters, reply tone, limits, plumbing — is in the Input tab and in Shape who you keep.

📤 Output

One dataset item per qualified lead — JSON, CSV, Excel or a webhook, best first.

Views: Shortlist, Leads (CRM import), Outreach (comment, DM) and Signals (why it qualified). OUTPUT holds the funnel, the charged events, yieldByQuery, aiTokens, history and sourceError / aiError — see Run outcomes.

🔁 Run it weekly

1. Actions → Schedule in the Console — daily with postedWithin: "24h", or weekly with "week".
2. Leave dedupeAcrossRuns: true — each run delivers and charges only new leads.
3. Add webhookUrl (Slack works as is) or notifyEmail — one digest per run: "7 new leads this run (2 hot) …".
4. pausePhrasesAfterEmptyRuns: 3 drops phrases that find nobody; outcomes (<dedupeKey> = meeting) reports meetings per phrase.

Weekly cost: $0.50 of search + ≈ 15 new leads × $0.009 + ≈ $0.09 tokens → ≈ $0.73. In Q4, planning season, add the planning_season pack.

🚨 Instant hot-lead alert

The first helpful reply wins the thread. Run a small monitor every hour and send only hot, fresh buyers to Slack:

Each post is alerted once; an hour without a new hot buyer costs only the search (8 phrases × 20 posts = $0.32).

🎯 Account intent monitor (ABM)

ABM teams know who they want to sell to — they need to know when. Put company URLs into watchCompanies (or people into watchAuthors) for LinkedIn account monitoring: every post where someone there asks for a recommendation.

🔌 Integrations, automation and API

• Slack, Zapier, Make, n8n — point webhookUrl at it (Slack works as is): {"event": "lead.qualified", "lead": {…}} per lead (or batches), then one {"event": "run.digest", …} per run; custom headers carry your API key.
• HubSpot / any CRM — map the Leads view or the webhook fields; dedupeKey is a stable external id.
• Google Sheets — one click from the Integrations tab.
• AI agents / MCP — every input is a plain typed field. The first call is the free trial: {"sourceMode": "list"}.

Already scraping LinkedIn? Qualify your LinkedIn scraper data

Ran harvestapi/linkedin-post-search or harvestapi/linkedin-profile-posts yourself? Add a webhook on its succeeded run that starts this Actor with sourceMode: "dataset", datasetId = that run's defaultDatasetId and enableAi: false — no second scrape, only the buyers in it are charged (rules only by default).

MCP prompt (server https://mcp.apify.com/?actors=rich_minds/linkedin-intent-leads): "Run rich_minds/linkedin-intent-leads with sourceMode actor, intentKeywords ["recommend a CRM"], offerDescription "

👥 Who is it for?

🧠 How the AI works

• One typed call per lead. The model (default Claude Haiku 4.5) sees the post, headline, rule cues and your offer.
• Grounded. Evidence must appear verbatim in the post, "high" urgency needs a deadline in the text, "not intent" caps the score at 20, and a reply with a pitch, a link or a placeholder is replaced by the template.
• Tokens only where they matter. Only the best rule-scored leads get the AI pass; "not intent" is dropped uncharged.

🔒 Data, compliance and limits

• Public data only, no cookies, no login. Public LinkedIn posts and their author cards, searched by harvestapi/linkedin-post-search — no private profiles, no e-mail guessing, no messages sent.
• Limits. Rule cues cover English, German, French, Spanish and Dutch (the AI reads any language); yield figures are estimates until the first public live run is published; post text is cut at 1,500 characters.
• Keep your outreach within LinkedIn's terms and the privacy law that applies to you (GDPR, CCPA); honour opt-outs with suppressionList.

❓ FAQ

Is this a LinkedIn post scraper? It is built on one: the LinkedIn post search runs on your account and finds the posts, then this Actor drops vendors, job ads and duplicates, folds each author into one lead and scores it against your offer. You pay only for the buyers.

Will two clients or briefs share one memory? No: the cross-run memory is kept per competitor set (competitorNames), so a second client with other competitors gets its own. Several clients with the same competitors: give each its own dedupeStoreName.

Which Apify plan do I need? The demo runs on any plan. A real search runs harvestapi/linkedin-post-search on your account, so it needs a plan that can run Store Actors (the free plan's $5 credit covers ≈ 15 first searches); so does Apify's AI model access — else set AI model access to My own API key.

Do I need LinkedIn cookies or a login? No. The search reads public posts without cookies or a login, so your LinkedIn account is never used or at risk.

How long does a run take? The demo takes seconds; a real run, the search + under a minute + at most 5 min of AI. From the API, wait ≥ 15 min.

I don't know which phrases buyers write. Where do I start? Describe what you sell: phrases are suggested from your offer. Or pick a phrase pack, or type only your competitors. OUTPUT.yieldByQuery shows which phrases find buyers.

What counts as a "qualified lead"? One LinkedIn author whose post shows a target intent, passed every free noise rule and your filters, and scored at least minScore. Several posts by one person are one lead.

What happens if nothing matches, or the search cannot run? Nothing matched: the run succeeds with 0 rows and charges nothing; the OUTPUT funnel shows where posts were dropped. A search that cannot run says why in sourceError — see Run outcomes.

Can I monitor a list of target accounts (ABM)? Yes: watchCompanies or watchAuthors, scheduled weekly.

Is this LinkedIn account monitoring for ABM intent? Yes, priced per person who posted a buying signal ($0.002–$0.009), not per account per year.

Does it catch recommendation requests and competitor switching? Both, plus buying, evaluating and hiring a service, with the triggering phrases quoted. competitorNames adds five switching searches per rival.

Is it for social selling or B2B lead qualification? Both: free rules, one lead per person, a score against your offer, and a comment + DM to post — empty for cold leads.

How are LinkedIn post leads different from other LinkedIn lead generation tools? Title lists leave you guessing who buys; each lead here is the author of a public post asking for it this week.

🧩 More Actors from the same developer

• Twitter Lead Generation — the same buyers, found on X.
• Reddit Brand Monitoring — threads naming you or a rival, best first.
• Instagram Creator Qualifier — creators your buyers follow, scored.
• AI Local Business Lead Gen — when your buyers are local businesses.
• Google Reviews AI Insights — a business's complaints, with an opener.
• Email List Cleaner — only the e-mails worth sending to.

🆘 Support

Something missing or wrong? Open an issue on the Actor's page. If it helped: A review on the Store helps others find it. Roadmap: "+1" commenters under hot posts, a work e-mail look-up, a native HubSpot / Pipedrive push.

📝 Changelog

• 0.5.5 (2026-10-16) — A search the plan cannot run ends SUCCEEDED with ok: false on schedules too; a source that renamed its fields fails at once, nothing charged.
• 0.5.4 (2026-10-04) — Pasted posts or a dataset ID pick the mode themselves.
• 0.6 — per-brief memory, a month of history on the first live run, sample AI verdicts, an AI trial on any plan, own-key rate-limit budgets.
• 0.1–0.5 — post search, noise rules, intent types, one lead per author, memory, AI replies, language packs, ABM, MCP / n8n recipes.