🔗 Backlink Building Agent: Discover, Contact & Draft Outreach

Overview

Backlink Building Agent runs the real backlink-outreach workflow end to end: give it a topic and your own site, and it (1) discovers real candidate pages worth pitching for a backlink, (2) extracts a real contact from each candidate site, and (3) drafts a genuinely personalized outreach email for you to review and send. It is a 3-stage agent pipeline, not a single-source scraper — most of the value is in wiring discovery, contact extraction, and AI drafting together so you end a run with something you can actually act on, not a spreadsheet of raw URLs.

This actor never sends email. It drafts outreach for a human to review and send themselves. Auto-sending is a much bigger scope with real consent and deliverability implications, and this actor deliberately stays out of that.

Billing is outcome-tiered, not per-row: you pay the most for a fully outreach-ready lead (discovered, contacted, and drafted), less for a lead that only made it partway, and nothing at all when a run finds nothing usable.

✅ No login required | ✅ Real contact extraction, not guessed emails | ✅ AI-personalized drafts, not templates | ✅ Pay per outcome | ✅ MCP-ready for AI agents

How it works

Stage 1 — Discover. For each topic you give it, the actor runs a set of standard backlink-prospecting search queries (topic + "write for us", "guest post guidelines", "helpful resources", etc., depending on your outreach goal) against Google Search, over Apify's purpose-built GOOGLE_SERP proxy tier. Results are filtered twice: a denylist drops reference sites, search engines and social platforms, and a relevance gate requires the actual footprint phrase to appear in the result before it counts as a candidate. A degraded or irrelevant search never produces fabricated prospects — it produces an honest zero.

Stage 2 — Extract a contact. Each qualified candidate domain is handed to the sibling actor website-contact-finder, called directly via the Apify API, which crawls the homepage plus contact/about/team pages for a real email, phone number, or social profile. This actor absorbs that call's own small per-domain cost rather than adding it as a separate line item on your bill — you pay one price for the outcome.

Stage 3 — Draft outreach. If a contact was found, the actor drafts a short, specific email with Gemini: it must reference something concrete from the target page (not generic flattery), explain why your content is relevant to their audience, and end with a single soft ask. A quality gate rejects anything that reads as templated (too short, leftover [placeholder] brackets, or no actual reference to the target page) — that draft is simply not produced rather than shipped as a weak one.

🧾 Input configuration

{
  "topics": ["sustainable jewelry"],
  "siteUrl": "https://example.com",
  "siteDescription": "We publish independently reviewed guides on lab-grown diamonds and ethical jewelry buying, including a widely-cited guide comparing certification standards.",
  "senderName": "Alex",
  "senderRole": "Content lead",
  "outreachGoal": "resource_link",
  "maxProspectsPerTopic": 5,
  "maxOutreachEmails": 5,
  "contactPagesPerSite": 5,
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["GOOGLE_SERP"] }
}

Field	Description
topics	Niche keywords to prospect for. Each one is searched independently.
siteUrl	Your site/page URL. Used for personalization and so the agent never pitches your own domain to itself.
siteDescription	1-3 sentences on what you're promoting. The single biggest lever for draft quality — be specific.
senderName / senderRole	Used in the drafted email's signature.
outreachGoal	guest_post, resource_link, or general_collaboration — changes both the search footprints and the email's ask.
maxProspectsPerTopic	How many footprint-matched candidates to discover per topic. Discovery is cheap (search only) and does not by itself drive cost.
maxOutreachEmails	The real spend/time cap. How many discovered prospects are carried through contact-extraction + drafting, across all topics combined, in this run.
contactPagesPerSite	Pages checked per candidate site for a contact (passed to website-contact-finder).
proxyConfiguration	Apify Proxy for the search stage (stage 1). GOOGLE_SERP — Apify's purpose-built Google Search proxy tier — is the default and what this actor is tuned for; residential proxy was tested and actually returns worse, query-degraded results for this specific target.

📤 Output format

Every discovered prospect produces one row, tagged with how far it got:

{
  "topic": "lab-grown diamonds",
  "outreach_goal": "resource_link",
  "pipeline_stage": "complete_outreach_lead",
  "target_domain": "example-jewelry-blog.com",
  "target_url": "https://example-jewelry-blog.com/resources",
  "target_title": "Ethical Jewelry Resources & Further Reading",
  "target_snippet": "A curated list of resources for buyers researching ethical and lab-grown jewelry options.",
  "footprint_matched": "helpful resources",
  "contact_email": "editor@example-jewelry-blog.com",
  "contact_phone": null,
  "contact_socials": null,
  "contact_source": "website-contact-finder",
  "outreach_subject": "A certification-standards resource for your ethical jewelry list",
  "outreach_body": "Hi there, I came across your curated resources page for ethical and lab-grown jewelry buyers...",
  "personalization_note": "References \"Ethical Jewelry Resources & Further Reading\" on example-jewelry-blog.com.",
  "scraped_at": "2026-08-02T10:00:00.000Z"
}

Field	Description
pipeline_stage	qualified_prospect (discovered, no contact found), contacted_prospect (contact found, draft failed/skipped), or complete_outreach_lead (all three stages succeeded)
target_domain / target_url	The candidate site and specific page that matched a prospecting footprint
footprint_matched	Which search pattern surfaced it (e.g. "write for us", "helpful resources")
contact_email / contact_phone / contact_socials	Best contact found, present from contacted_prospect onward
outreach_subject / outreach_body	The drafted email — present only at complete_outreach_lead. Review before sending; this actor never sends it for you.
personalization_note	What the draft specifically referenced, for your own QA pass

💰 Pricing: outcome-tiered, not per-row

Backlink outreach tools on Apify are typically priced per finished outcome rather than per raw scraped row, because a raw URL and a send-ready lead are very different things to build and very different things to be worth. This actor follows the same logic with three tiers instead of one flat price, so you are never charged the full price for a partial result:

Outcome	What happened	Why this tier
Complete outreach lead	Discovered + contact extracted + email drafted	The full product: something you can review and send today
Contacted prospect	Discovered + contact extracted, drafting failed or was skipped	You still got a real, verified contact to reach out to
Qualified prospect	A real footprint-matched candidate site, no public contact found	Still a genuine, filtered lead worth manual follow-up
(nothing usable)	No candidates matched, or nothing survived the relevance gate	Free — nothing is ever charged for an empty or degraded run

Pricing is pay-per-event with the standard tiered discount for higher Apify plans. Check the Actor's pricing tab for current rates.

💼 Common use cases

SEO backlink outreach Turn a target keyword or niche into a working outreach list without hiring an outreach VA to do the search-and-scrape legwork by hand.

PR & media outreach Point it at a topic your story/data is relevant to and get a starting list of relevant sites plus a drafted pitch angle for each.

Content promotion After publishing a new guide or piece of original research, use it to find resource pages and roundups it's a genuine fit for.

Agency link-building at scale Run it per client niche to produce a first-pass prospect + draft list, then have a human do final QA and send.

Honest limits (read before you run a big batch)

• Discovery quality depends on how much "write for us" / "resources" content already exists for your niche. Broad, well-covered niches (SaaS, finance, health) turn up plenty of candidates. Narrow or under-written niches can legitimately return few or none — that is an honest signal about the niche, not a broken run. See BACKLINK-OUTREACH.md in our own build notes: even hand-researching our own niche found genuinely neutral targets are rare.
• Contact coverage is never 100%. Not every site publishes a public email; some only have a contact form, which this actor cannot fill out for you.
• AI drafts still need a human read before sending. The quality gate rejects obviously generic output, but it cannot verify facts about the target site beyond what the search snippet says — skim each draft before you send it.
• This actor never sends anything. No email account is connected, and there is no send feature to turn on.

FAQ

Does this actually send the outreach emails? No. It drafts a subject and body for each complete lead. You copy it into your own email client (or a mail-merge tool) and send it yourself, after your own review.

How is a "contact" verified? It is extracted from a real, public page on the target site (not looked up in a third-party database or guessed from a name pattern) by the sibling website-contact-finder actor.

What if my niche returns very few results? That's an honest outcome, not a bug — see "Honest limits" above. Try a broader topic phrasing, or a different outreachGoal (resource-link footprints and guest-post footprints surface different sites).

Can I use it in an AI agent? Yes. It's exposed as an MCP tool. See below.

Use in Claude, ChatGPT & any MCP agent

https://mcp.apify.com/?tools=themineworks/backlink-building-agent

Or call it programmatically with the Apify client:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });

const run = await client.actor('themineworks/backlink-building-agent').call({
  topics: ['sustainable jewelry'],
  siteUrl: 'https://example.com',
  siteDescription: 'We publish independently reviewed guides on ethical jewelry buying.',
  senderName: 'Alex',
  outreachGoal: 'resource_link',
  maxOutreachEmails: 5,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

🛠️ Complete your outreach pipeline

• Website Contact Scraper: the sibling actor this pipeline calls for stage 2 — use it standalone if you already have your own list of target domains.
• B2B Leads Finder: find named decision-makers (not just a general inbox) at any company.
• Google News Scraper: monitor press coverage of your topic for fresh PR/media outreach angles.

Questions or need a custom field set? Reach out through the Apify profile.