How We Built an AI Email Finder with 5x Better Match Rates

• Scrapers that return generic emails (info@, hello@, etc.)

• APIs that give outdated, low-confidence guesses

• Finds real, personalized emails

• Gives context ("this belongs to the creator")

• Links to social handles when possible

• Achieves 5x higher match rates than traditional tools

🤖 Our Stack: AI + APIs + Orchestration

• Gemini 2.5 Pro: The brain of the operation, understanding website content like a human.

• Hunter API: For verifying and enriching contact data.

• DataForSEO: For fetching the SERP data that feeds our Keyword Analyzer.

• Trigger.dev: The conductor, orchestrating all the background tasks seamlessly.

• PostgreSQL: Our trusty database for caching results, scoring contacts, and preventing duplicates.

⚡️ The Workflow: From Domain to Contact

1. Intelligent Assessment

// Pseudocode: Decide what needs to be filled
const { needsAIFill, needsHunterFill, needsContactFill } = getDomainMissingFields(domain)

2. Parallel Processing

await Promise.allSettled([
  needsAIFill && fillWebsiteDetails(domain),
  needsHunterFill && getHunterDomainSearch(domain),
  needsContactFill && findEmailsForDomain(domain),
].filter(Boolean))

3. AI Deep Dive

// Pseudocode: AI parses HTML and returns contacts with context
const aiContacts = await gemini.extractContactsFromHtml(html, { domain })
/*
  [
    {
      email: "[email protected]",
      type: "personal",
      note: "Found in footer, likely belongs to the creator 'Alex'",
      twitter: "@alexdev"
    }
  ]
*/

4. Smart Merging

for (const contact of aiContacts) {
  await db.domainContact.upsert({
    where: { value: contact.email },
    create: { ...contact, domain },
    update: { ...contact, updatedAt: new Date() }
  })
}

5. Contextual Notes

🕵️️ Resilience & Automation: Trigger.dev in Action

• Efficient Batching: When multiple domains need checking (like after analyzing a keyword’s search results), Trigger.dev figures out which ones actually require updates and queues them up in batches.

• Smart Task Management: It schedules background jobs for each domain, managing parallel data fetching and AI analysis.

• Built-in Resilience: Web scraping and API calls can be flaky. Trigger.dev automatically retries failed steps. We even have a dedicated process that periodically sweeps for domains that failed previously due to temporary issues (like API quotas) and gives them another shot.

// Pseudocode: Batch trigger for domains needing refresh
await fillDomainTask.batchTrigger(
  domainsToFill.map(domain => ({ payload: { domain } }))
)

• Seamless Background Processing: All this happens reliably behind the scenes, ensuring our data stays fresh without interrupting your workflow in LinkDR.

🏆 Real-World Example: How We Find the Right Contact

1. SERP Fetch: We pull the top 100 results from Google.

2. Domain Check: For each domain, we check if we already have up-to-date contact info.

3. Background Fill: For missing/stale domains, we trigger the AI/Hunter pipeline in the background.

4. AI Reasoning: Gemini 2.5 Pro might find a line like “Built by Alex” in the footer, spot an email, and link it to a Twitter handle.

5. Contact Merge: The system upserts this as a "personal" contact, with a note: "Found in footer, likely belongs to the creator ‘Alex’ (@alexdev)."

6. Instant UI: You see which domains have contacts right away. As background jobs finish, new contacts appear—no need to refresh.

✅ The Results Speak for Themselves

• 5x Higher Match Rate: Finding valid, personal emails far more often than standalone tools.

• Actionable Context: Users love the AI notes (“This email likely belongs to the founder, mentioned on the About page”).

• Lightning-Fast Lookups: Cached results mean common domains return contact info instantly.

• Improved Outreach: Lower bounce rates and better connection potential.

• Smarter Automation: Providing high-quality inputs for Patricia AI, our automated backlink agent.

• Privacy First: Sensitive details like full emails and names are automatically obscured for non-admin users viewing domain info within LinkDR. We use these internally when placing links for you. Check out pricing for more info

📄 See It In Action: Try the Keyword Analyzer

1. Enter a keyword you want to rank for.

2. LinkDR fetches the latest search results.

3. We instantly compare the ranking domains against our internal database.

4. The results page immediately highlights which sites already have contact information available in LinkDR (with sensitive details kept private for non-admins).

5. Behind the scenes, we automatically queue up any domains needing a data refresh, ensuring our intelligence stays up-to-date.

6. You can also manually trigger a refresh for any specific domain right from the results.

7. The handy “Copy SERP Analysis” button includes the best available contact details, ready for your outreach workflow.